HK1232961B - Pattern drawing device, pattern drawing method, device manufacturing method, laser light source device, beam scanning device, and beam scanning method - Google Patents

Description

Pattern drawing device, pattern drawing method, device manufacturing method, laser light source device, beam scanning device and method

Technical Field

The present invention relates to a beam scanning device and a beam scanning method for scanning a point light of a light beam irradiated on an irradiated object, a pattern drawing device and a pattern drawing method for scanning a point light to draw a predetermined pattern on the irradiated object, a device manufacturing method using the pattern drawing method, and a laser light source device used in the pattern drawing device and the beam scanning device.

Background Art

As disclosed in Japanese Patent Application Laid-Open Nos. 61-134724 and 2001-133710, there are known laser irradiation devices and laser drawing devices that split a laser beam from a single laser oscillator (laser beam source) into two by a half-mirror, and then cause each of the split laser beams to enter two polygon mirrors (rotating polygon mirrors), thereby causing the two laser beams to scan the drawing object. Japanese Patent Application Laid-Open No. 2001-133710 also discloses a method in which the two split laser beams incident on the two polygon mirrors are modulated by passing each of the split laser beams through an AOM (acousto-optic modulator), where the AOM is turned on/off in response to drawing data.

Summary of the Invention

However, in beam scanning using a polygon mirror, there are periods during the polygon mirror's rotation during which the incident laser beam cannot be effectively reflected toward the object being drawn, depending on the number of reflective surfaces of the polygon mirror, the incident conditions of the optical system (e.g., the fθ lens) after the polygon mirror, and other factors. Therefore, even if the laser beam is split into two by a half-mirror and incident on two polygon mirrors as is conventionally done, there are periods during which the laser beam cannot be effectively irradiated onto the object being drawn, i.e., non-drawing periods, and the laser beam from the light source cannot be effectively utilized.

The first embodiment of the present invention is a pattern drawing device that draws a specified pattern on an irradiated object by scanning points of a laser, and comprises: a light source device that emits the above-mentioned laser; a plurality of drawing units that are used to generate the above-mentioned scanning points by incidenting the above-mentioned laser, and include a light scanning component and an optical lens system for scanning with the above-mentioned laser, and are configured to scan the above-mentioned scanning points on different areas on the above-mentioned irradiated object; and a plurality of selection optical elements that are arranged in series along the direction of travel of the above-mentioned laser from the above-mentioned light source device in order to switch whether the above-mentioned laser from the above-mentioned light source device is incident on the selected above-mentioned drawing unit among the above-mentioned multiple drawing units.

The second embodiment of the present invention is a pattern drawing device that draws a prescribed pattern on an irradiated object by means of scanning points of a laser, and comprises: a light source device that emits the laser; a plurality of drawing units that include a light scanning component and an optical lens system for scanning with the laser in order to generate the scanning points by incident on the laser, and are arranged so that the scanning points scan different areas on the irradiated object; a plurality of selection optical elements that are arranged in series along the direction of travel of the laser from the light source device in order to selectively allow the laser from the light source device to be incident on the plurality of drawing units; and a drawing light modulator that modulates the intensity of the laser incident on the plurality of selection optical elements based on the drawing data of each of the plurality of drawing units that specify the pattern to be drawn on the irradiated object by the scanning points.

The third scheme of the present invention comprises: a pulse light source device, which generates a pulsed light beam with an adjustable oscillation period; a first drawing unit, which projects the light beam from the above-mentioned pulse light source device as a point light onto the irradiated object, and deflects the above-mentioned light beam in a manner that repeats a prescribed period during the projection period and the non-projection period of the point light onto the above-mentioned irradiated object, and scans the above-mentioned point light along a first drawing line on the above-mentioned irradiated object during the above-mentioned projection period; and a second drawing unit, which projects the light beam from the above-mentioned pulse light source device as a point light onto the above-mentioned irradiated object, and deflects the above-mentioned light beam in a manner that repeats a prescribed period during the projection period and the non-projection period, and scans the above-mentioned point light along a second drawing line on the above-mentioned irradiated object different from the above-mentioned first drawing line during the above-mentioned projection period. The drawing line is scanned; a first control system, which synchronously controls the first drawing unit and the second drawing unit so that the projection period of the first drawing unit corresponds to the non-projection period of the second drawing unit, and the projection period of the second drawing unit corresponds to the non-projection period of the first drawing unit; and a second control system, which controls the pulse light source device so that during the projection period in the first drawing unit, the oscillation of the light beam is controlled based on the first drawing information of the pattern to be drawn by the first drawing line, and during the projection period in the second drawing unit, the oscillation of the light beam is controlled based on the second drawing information of the pattern to be drawn by the second drawing line.

The fourth scheme of the present invention is a pattern drawing device, which modulates the intensity of the point light of the ultraviolet laser converged on the irradiated object according to the drawing data, and scans the above-mentioned point light and the above-mentioned irradiated object relative to each other, thereby drawing a pattern on the above-mentioned irradiated object, comprising: a laser light source device, which includes a light source unit that generates seed light as the source of the above-mentioned ultraviolet laser, an optical amplifier that receives the above-mentioned seed light and amplifies it, and a wavelength conversion optical element that generates the above-mentioned ultraviolet laser from the amplified above-mentioned seed light; and a drawing modulation device, which modulates the intensity of the above-mentioned seed light generated from the above-mentioned light source unit according to the above-mentioned drawing data in order to modulate the intensity of the above-mentioned point light.

The fifth scheme of the present invention is a pattern drawing method, which modulates the intensity of the point light of the ultraviolet laser converged on the irradiated object according to the drawing data, and scans the above-mentioned point light and the above-mentioned irradiated object relative to each other, thereby drawing a pattern on the above-mentioned irradiated object, including: a conversion step, amplifying the seed light serving as the source of the above-mentioned ultraviolet laser by an optical amplifier, and converting the amplified above-mentioned seed light into the above-mentioned ultraviolet laser by a wavelength conversion optical element; and a modulation step, in order to modulate the intensity of the above-mentioned point light, modulating the intensity of the above-mentioned seed light incident to the above-mentioned optical amplifier according to the above-mentioned drawing data.

The sixth embodiment of the present invention is a device manufacturing method, comprising: while moving a photosensitive substrate prepared as the above-mentioned irradiated body along the first direction, drawing a device pattern on the photosensitive layer of the above-mentioned substrate by the pattern drawing method of the above-mentioned fifth embodiment; and selectively forming a specified pattern material according to the difference between the irradiated part and the non-irradiated part of the above-mentioned point light of the above-mentioned photosensitive layer.

The seventh scheme of the present invention is a laser light source device, which is connected to a device that draws a pattern by converging point light on an irradiated object, and emits a light beam that becomes the above-mentioned point light, comprising: a first semiconductor light source, which responds to a clock pulse of a specified period to generate a first pulse light with a luminous time that is short relative to the above-mentioned specified period and a high peak intensity, which rises and falls sharply; a second semiconductor light source, which responds to the above-mentioned clock pulse to generate a wide second pulse light with a luminous time that is shorter than the above-mentioned specified period and longer than the luminous time of the above-mentioned first pulse light and a low peak intensity; an optical fiber amplifier, which is incident with the above-mentioned first pulse light or the above-mentioned second pulse light; and a switching component, which performs optical switching based on the input of pattern information to be drawn, so that the above-mentioned first pulse light is incident on the above-mentioned optical fiber amplifier when the above-mentioned point light is projected onto the above-mentioned irradiated object, and the above-mentioned second pulse light is incident on the above-mentioned optical fiber amplifier when the above-mentioned point light is not projected onto the above-mentioned irradiated object.

The eighth scheme of the present invention is a light beam scanning device, which is provided with a plurality of scanning units in a prescribed positional relationship, the scanning unit comprising a rotating polygonal mirror for repeatedly deflecting a light beam from a light source device, and a projection optical system for incidenting the deflected light beam and converging the light beam into a point light for performing one-dimensional scanning on an irradiated object, the light beam scanning device comprising: a light beam switching component for switching the optical path of the light beam in such a manner that the light beam from the light source device is incident on one of the plurality of scanning units that performs one-dimensional scanning of the point light; and a light beam switching control unit for controlling the light beam switching component so that the deflection of the light beam achieved by the rotating polygonal mirror of the scanning unit is repeated at every at least one reflecting surface of the rotating polygonal mirror, so that each of the plurality of scanning units performs one-dimensional scanning of the point light in sequence.

The ninth scheme of the present invention is a light beam scanning device, comprising a plurality of scanning modules each having a plurality of scanning units arranged in a prescribed positional relationship, the scanning unit comprising a rotating polygonal mirror that rotates at a certain rotational speed in order to repeatedly deflect a light beam from a light source device, and a projection optical system that incidents the deflected light beam and converges the incident deflected light beam into a point light for one-dimensional scanning on an irradiated object, the light beam scanning device comprising: a light beam switching component that switches the optical path of the light beam in such a manner that the light beam from the light source device is incident on the scanning unit among the plurality of scanning units that performs one-dimensional scanning of the point light; and a light beam switching control unit that controls the light beam switching component so that the deflection of the light beam achieved by the rotating polygonal mirror of each of the scanning units is switched to either a first state that is repeated on each continuous reflection surface of the rotating polygonal mirror or a second state that is repeated on at least one reflection surface of the rotating polygonal mirror, so that each of the plurality of scanning units performs one-dimensional scanning of the point light in sequence.

The tenth scheme of the present invention is a light beam scanning method, in which a plurality of scanning units are arranged in a prescribed positional relationship to perform light beam scanning on an irradiated object, the scanning unit having a projection optical system for incident light beams repeatedly deflected by a rotating polygonal mirror and converging the light beams into point light for one-dimensional scanning on the irradiated object, the light beam scanning method comprising: synchronously rotating the plurality of rotating polygonal mirrors in such a manner that the rotation angle positions of the respective rotating polygonal mirrors of the plurality of scanning units are in a prescribed phase relationship with each other; and switching the scanning unit on which the light beam is incident in such a manner that the one-dimensional scanning of the point light performed by each of the plurality of scanning units is performed sequentially and the deflection of the light beam realized by the rotating polygonal mirror is repeatedly performed for at least one reflecting surface of the rotating polygonal mirror.

The eleventh embodiment of the present invention is a light beam scanning method, in which a light beam is scanned on an irradiated object by a light beam scanning device having a plurality of scanning units arranged in a predetermined positional relationship, the plurality of scanning units having a projection optical system for incident light beams repeatedly deflected by a rotating polygonal mirror rotating at a certain rotational speed and converging the light beams into point lights for one-dimensional scanning on the irradiated object, the light beam scanning method comprising: synchronously rotating the plurality of rotating polygonal mirrors in such a manner that the rotational angle positions of the respective rotating polygonal mirrors of the plurality of scanning units are in a predetermined phase relationship with each other; a first scanning step, switching the scanning units on which the light beams are incident in such a manner that the deflection of the light beams achieved by the rotating polygonal mirrors is repeatedly performed on each successive reflecting surface of the rotating polygonal mirror, whereby the plurality of scanning units each perform one-dimensional scanning of the point lights in sequence; a second scanning step A scanning process is to switch the scanning unit on which the light beam is incident so that the deflection of the light beam achieved by the rotating polygonal mirror is repeated at every at least one reflecting surface of the rotating polygonal mirror, thereby each of the plurality of scanning units sequentially performs one-dimensional scanning of the point light; a switching process is to switch the first scanning process and the second scanning process.

The twelfth scheme of the present invention is a pattern drawing method, which uses a drawing device, which arranges a plurality of scanning units that cause the point light of the light beam from the light source device to perform main scanning along the drawing line to be joined on the substrate along the main scanning direction of the above-mentioned drawing line according to the pattern drawn by each drawing line, so that the above-mentioned plurality of scanning units and the above-mentioned substrate are relatively moved in a sub-scanning direction intersecting with the above-mentioned main scanning direction. The pattern drawing method includes: selecting a specific scanning unit among the above-mentioned plurality of scanning units corresponding to the width of the above-mentioned substrate in the above-mentioned main scanning direction, or the width or position of the exposure area of the above-mentioned substrate to be drawn in the above-mentioned main scanning direction; and after modulating the intensity of the above-mentioned light beam based on the pattern data to be drawn by each of the above-mentioned specific scanning units, selectively supplying it to each of the above-mentioned specific scanning units in sequence via a beam distribution unit that sends the above-mentioned light beam from the above-mentioned light source device.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagram showing a schematic configuration of a device manufacturing system including an exposure apparatus for performing an exposure process on a substrate according to a first embodiment.

FIG. 2 is a diagram showing a support frame that supports the drawing head and the rotating drum shown in FIG. 1 .

FIG. 3 is a diagram showing the structure of the drawing head of FIG. 1 .

FIG. 4 is a detailed structural diagram of the light introducing optical system shown in FIG. 3 .

FIG. 5 is a diagram showing a drawing line obtained by scanning a spot light by each scanning unit shown in FIG. 3 .

FIG. 6 is a diagram showing the relationship between the polygon mirrors of the scanning units shown in FIG. 3 and the scanning direction of the drawing line.

FIG. 7 is a diagram for explaining the rotation angle of the polygon mirror shown in FIG. 3 so that the reflecting surface of the polygon mirror can deflect (reflect) the laser light so as to be incident on the f-θ lens.

FIG. 8 is a diagram schematically illustrating an optical path between the light introducing optical system and a plurality of scanning units shown in FIG. 3 .

FIG. 9 is a diagram showing the configuration of an image drawing head in a modified example of the first embodiment.

FIG10 is a detailed structural diagram of the light introducing optical system shown in FIG9.

FIG. 11 is a diagram showing the structure of an image drawing head according to a second embodiment.

FIG. 12 is a diagram showing the light introducing optical system shown in FIG. 11 .

FIG. 13 is a diagram schematically illustrating an optical path between the light introducing optical system shown in FIG. 12 and a plurality of scanning units.

FIG. 14 is a block diagram showing an example of a control circuit for rotationally driving each polygon mirror of the plurality of scanning units shown in FIG. 13 .

FIG15 is a timing chart showing an example of the operation of the control circuit shown in FIG14 .

FIG. 16 is a block diagram showing an example of a circuit for generating drawing bit string data to be supplied to the drawing optical element shown in FIG. 11 to FIG. 13 .

FIG. 17 is a diagram showing a configuration of a light source device according to a modified example of the second embodiment.

FIG. 18 is a block diagram showing the configuration of a control unit for drawing control according to the third embodiment.

FIG. 19 is a timing chart showing signal states of various parts of the control unit of FIG. 18 and an oscillation state of laser light when the control unit draws a pattern.

FIG. 20 is a timing chart showing a clock signal for pulse light oscillation generated by the control circuit of the light source device of FIG. 17 .

FIG. 21 is a timing chart illustrating how the clock signal in FIG. 20 is corrected in order to correct the drawing magnification.

FIG. 22 is a diagram illustrating a method of correcting the drawing magnification in one drawing line (scanning line).

FIG. 23 is a diagram showing a schematic configuration of a device manufacturing system including an exposure apparatus for performing an exposure process on a substrate according to a fourth embodiment.

FIG. 24 is a detailed view of the rotating drum of FIG. 23 on which the substrate is wound.

FIG. 25 is a diagram showing a drawing line of spot light and an alignment mark formed on a substrate.

FIG26 is a structural diagram of a beam switching component.

27A is a diagram showing the switching of the optical paths of light beams by the selection optical element as viewed from the +Z direction side, and FIG. 27B is a diagram showing the switching of the optical paths of light beams by the selection optical element as viewed from the −Y direction side.

FIG28 is a diagram showing the optical structure of the scanning unit.

FIG29 is a diagram showing the structure of an origin sensor provided around the polygon mirror of FIG28.

FIG30 is a timing chart showing the relationship between the generation timing of the origin signal and the drawing start timing.

FIG31 is a block diagram of a sub-origin generating circuit for generating a sub-origin signal whose generation timing is delayed by a predetermined time by thinning out the origin signal.

FIG32 is a timing chart showing a sub-origin signal generated by the sub-origin generating circuit of FIG31 .

FIG33 is a block diagram showing the electrical structure of the exposure device.

FIG34 is a timing chart showing the timing of outputting the origin signal, the sub-origin signal, and the serial data.

FIG35 is a diagram showing the configuration of the drawing data output control unit shown in FIG33 .

FIG36 is a structural diagram of a light beam switching component according to the fifth embodiment.

FIG37 is a diagram showing an optical path when the position of the configuration switching member in FIG36 is the first position.

FIG38 is a diagram showing the configuration of a light beam switching control unit in the fifth embodiment.

FIG39 is a diagram showing the structure of the logic circuit of FIG38.

FIG40 is a timing diagram illustrating the operation of the logic circuit of FIG39.

FIG41 is a structural diagram of a light beam switching component according to the sixth embodiment.

FIG42 is a diagram showing a configuration in which the arrangement of the optical element for selection (acousto-optic modulation element) in the sixth embodiment is rotated by 90 degrees.

FIG. 43 is a diagram showing the relationship between the conveyance form of the substrate and the arrangement of the drawing lines according to Modification 3. FIG.

FIG44 is a diagram showing the configuration of a driver circuit for a selection optical element (acousto-optic modulation element) according to Modification 5.

FIG45 is a diagram showing a modified example of the driver circuit in FIG44.

DETAILED DESCRIPTION

Hereinafter, preferred embodiments of the pattern drawing device, pattern drawing method, beam scanning device, beam scanning method, device manufacturing method and laser light source device according to the present invention will be given and described in detail with reference to the accompanying drawings. In addition, the present invention is not limited to these embodiments, but also includes embodiments obtained by applying various changes or improvements. That is, the structural elements described below include elements that can be easily conceived by those skilled in the art and substantially the same elements, and the structural elements described below can be appropriately combined. In addition, various omissions, replacements or changes of structural elements can be made without departing from the scope of the present invention.

[First embodiment]

FIG1 is a diagram schematically illustrating the configuration of a device manufacturing system 10 including an exposure apparatus EX for performing an exposure process on a substrate (irradiated object) FS according to a first embodiment. In the following description, unless otherwise specified, an XYZ orthogonal coordinate system is assumed, with the direction of gravity being the Z direction, and the X, Y, and Z directions are described according to the arrows shown in the figures.

The device manufacturing system 10 is a manufacturing system that is constructed with a production line for manufacturing flexible displays, flexible wiring, flexible sensors, etc. as electronic devices. The following description is based on the premise that the electronic device is a flexible display. As flexible displays, there are, for example, organic EL displays, liquid crystal displays, etc. The device manufacturing system 10 has a so-called roll-to-roll structure, that is, a substrate FS is fed out from an unillustrated supply roller on which a flexible sheet-like substrate (sheet substrate) FS is rolled up, and after various processes are continuously performed on the fed substrate FS, the substrate FS after various processes is taken up by an unillustrated recovery roller. The substrate FS has a strip shape with the moving direction of the substrate FS being the long side direction (long side) and the width direction being the short side direction (short side). The substrate FS fed out from the above-mentioned supply roller passes through the processing device PR1, the exposure device (pattern drawing device, light beam scanning device) EX and the processing device PR2 in sequence and undergoes various processes, and is then taken up by the above-mentioned recovery roller.

The X direction is the direction (conveying direction) from the processing unit PR1 to the processing unit PR2 via the exposure unit EX in the horizontal plane. The Y direction is the direction perpendicular to the X direction in the horizontal plane and is the width direction (short side direction) of the substrate FS. The Z direction is the direction perpendicular to the X and Y directions (upward direction) and is parallel to the direction of gravity.

The substrate FS uses, for example, a resin film, or a foil made of a metal or alloy such as stainless steel. As the material of the resin film, materials including at least one of polyethylene resin, polypropylene resin, polyester resin, ethylene-vinyl alcohol copolymer resin, polyvinyl chloride resin, cellulose resin, polyamide resin, polyimide resin, polycarbonate resin, polystyrene resin and vinyl acetate resin can be used. In addition, the thickness and rigidity (Young's modulus) of the substrate FS only need to be within a range that does not produce creases or irreversible wrinkles due to bending on the substrate FS when passing through the conveying path of the exposure device EX. As the base material of the substrate FS, films such as PET (polyethylene terephthalate) and PEN (polyethylene naphthalate) with a thickness of about 25μm to 200μm are typical of preferred sheet substrates.

Since the substrate FS is subject to heat during the various processes performed by the processing unit PR1, the exposure unit EX, and the processing unit PR2, it is preferable to select a material for the substrate FS that does not have a significantly high thermal expansion coefficient. For example, the thermal expansion coefficient can be suppressed by mixing an inorganic filler into the resin film. The inorganic filler may be, for example, titanium oxide, zinc oxide, aluminum oxide, or silicon oxide. Alternatively, the substrate FS may be a single layer of ultra-thin glass having a thickness of approximately 100 μm, manufactured by a float process or the like, or a laminate formed by laminating the aforementioned resin film, foil, etc. to the ultra-thin glass.

In addition, the flexibility of the substrate FS refers to the property of the substrate FS that it can bend without being sheared or broken even if a force of the degree of its own weight is applied to the substrate FS. In addition, the property of bending due to the force of the degree of its own weight is also included in flexibility. In addition, the degree of flexibility will change depending on the material, size, thickness, layer structure formed on the substrate FS, temperature, humidity and other environmental factors of the substrate FS. In either case, as long as the substrate FS is smoothly conveyed without being buckled and causing creases or damage (breaking or cracking) when the substrate FS is correctly wound on the various conveying rollers, rotating drums and other conveying direction conversion components provided on the conveying path within the device manufacturing system 10 of the first embodiment, it can be said to be within the flexibility range.

The processing unit PR1 performs pre-processing on the substrate FS to be exposed by the exposure unit EX. The processing unit PR1 transports the pre-processed substrate FS to the exposure unit EX. Through this pre-processing, the substrate FS transported to the exposure unit EX becomes a substrate (photosensitive substrate) with a photosensitive functional layer (photosensitive layer, photosensitive layer) formed on its surface.

The photosensitive functional layer is applied as a solution on the substrate FS and becomes a layer (film) by drying. The photosensitive functional layer is typically a photoresist (liquid or dry film), but as a material that does not require development treatment, there are photosensitive silane coupling agents (SAM) whose lyophilicity of the part exposed to ultraviolet light is modified, or photosensitive reducing agents that show plating reducing groups on the part exposed to ultraviolet light. In the case of using a photosensitive silane coupling agent as the photosensitive functional layer, the pattern part on the substrate FS that is exposed to ultraviolet light is modified from lyophobic to lyophilic. Therefore, by selectively applying a conductive ink (ink containing conductive nanoparticles of silver, copper, etc.) or a liquid containing a semiconductor material on the part that becomes lyophilic, it is possible to form a pattern layer that constitutes an electrode, a semiconductor, an insulation, or a wiring or electrode for connection of a thin film transistor (TFT). In the case of using a photosensitive reducing agent as the photosensitive functional layer, a plating reducing group is shown on the pattern part on the substrate that is exposed to ultraviolet light. Therefore, after exposure, the substrate FS is immediately immersed in a plating solution containing palladium ions for a certain period of time to form (precipitate) a pattern layer based on palladium. This plating process is an additive process. However, in addition to this, when etching is a subtractive process, the substrate FS fed to the exposure apparatus EX can be made of a base material of PET or PEN, and a metallic thin film such as aluminum (Al) or copper (Cu) can be vapor-deposited entirely or selectively on its surface, and a photoresist layer can be stacked thereon.

In this first embodiment, the exposure device EX is an exposure device of a direct drawing method that does not use a mask, that is, an exposure device of a so-called raster scan method. The exposure device EX irradiates the irradiated surface (photosensitive surface) of the substrate FS supplied from the processing device PR1 with a light pattern corresponding to a prescribed pattern for electronic devices, circuits, or wiring for a display. Although it will be described in detail later, the exposure device EX transports the substrate FS in the +X direction (sub-scanning direction) while causing the point light SP of the exposure light beam (laser, irradiation light) LB to scan one dimensionally on the substrate FS (the irradiated surface of the substrate FS) along a prescribed scanning direction (Y direction), while modulating (on/off) the intensity of the point light SP at high speed according to the pattern data (drawing data, drawing information). As a result, a light pattern corresponding to a prescribed pattern of electronic devices, circuits, or wiring is drawn and exposed on the surface (photosensitive surface) of the substrate FS that is the irradiated surface. That is, through the sub-scanning of the substrate FS and the main scanning of the point light SP, the point light SP is relatively scanned two-dimensionally on the irradiated surface of the substrate FS, thereby drawing and exposing a prescribed pattern on the substrate FS. In addition, since the substrate FS is transported along the transport direction (+X direction), a plurality of exposure areas W to which the pattern is exposed by the exposure device EX are provided at prescribed intervals along the long side direction of the substrate FS (refer to Figure 5). Since electronic devices are formed in this exposure area W, the exposure area W is also the electronic device formation area. In addition, the electronic device is formed by overlapping a plurality of pattern layers (layers on which patterns are formed), so the pattern corresponding to each layer can also be exposed by the exposure device EX.

The processing device PR2 performs subsequent process processing (such as plating processing or development/etching processing) on the substrate FS subjected to the exposure process by the exposure device EX. Through the subsequent process processing, a pattern layer of the device is formed on the substrate FS.

As described above, since electronic devices are constructed by overlapping multiple pattern layers, a pattern layer is generated through at least each process of the device manufacturing system 10. Therefore, in order to generate electronic devices, each process of the device manufacturing system 10 as shown in Figure 1 must be performed at least twice. To this end, by installing a recovery roller wound with the substrate FS as a supply roller to other device manufacturing systems 10, pattern layers can be stacked. Electronic devices are formed by repeating such actions. Therefore, the processed substrate FS becomes a state in which multiple electronic devices (exposure areas W) are connected at specified intervals along the long side direction of the substrate FS. In other words, the substrate FS is a substrate for simultaneous processing of multiple pieces.

The recovery roller that recovers the substrate FS formed in a state where the electronic components are connected can also be installed on a cutting device not shown. The cutting device equipped with the recovery roller divides (cuts) the processed substrate FS according to each electronic component (electronic component formation area W), thereby making it into multiple electronic components. Regarding the size of the substrate FS, for example, the size in the width direction (the direction of the short side) is about 10 cm to 2 m, and the size in the length direction (the direction of the long side) is more than 10 m. In addition, the size of the substrate FS is not limited to the above-mentioned size.

Next, the exposure device EX will be described in detail. The exposure device EX is housed in a temperature control chamber ECV. The temperature control chamber ECV suppresses temperature-induced shape changes of the substrate FS being transported inside by maintaining the interior at a specified temperature. The temperature control chamber ECV is arranged on the setting surface E of the production plant via passive or active vibration isolation units SU1 and SU2. The vibration isolation units SU1 and SU2 reduce vibrations from the setting surface E. The setting surface E can be the floor of the factory itself or a surface on a setting table (pedestal) set on the ground to form a horizontal surface. The exposure device EX includes a substrate transport mechanism 12, a light source device (pulse light source device, laser light source device) 14, a drawing head 16, and a control device 18.

The substrate transport mechanism 12 transports the substrate FS, transferred from the processing unit PR1, within the exposure unit EX at a predetermined speed and then delivers it to the processing unit PR2 at a predetermined speed. The substrate transport mechanism 12 defines the transport path of the substrate FS within the exposure unit EX. The substrate transport mechanism 12 comprises, in order from the upstream side (-X direction) in the transport direction of the substrate FS, an edge position controller EPC, a drive roller R1, a tension adjustment roller RT1, a rotating drum (cylindrical roller) DR, a tension adjustment roller RT2, a drive roller R2, and a drive roller R3.

The edge position controller EPC adjusts the position of the substrate FS conveyed from the processing device PR1 in the width direction (Y direction, short side direction of the substrate FS). That is, the edge position controller EPC moves the substrate FS in the width direction so that the position of the end (edge) in the width direction of the substrate FS conveyed under a specified tension converges within a range of ±10 μm to several tens of μm relative to the target position (allowable range). The edge position controller EPC has a roller for the substrate FS to hang on, and an edge sensor (end detection unit) (not shown) for detecting the position of the end (edge) in the width direction of the substrate FS. Based on the detection signal detected by the edge sensor, the above-mentioned roller of the edge position controller EPC is moved in the Y direction to adjust the position of the substrate FS in the width direction. The drive roller R1 rotates while holding the front and back surfaces of the substrate FS conveyed from the edge position controller EPC, conveying the substrate FS toward the rotating drum DR. In addition, the edge position controller EPC can also appropriately adjust the position of the substrate FS in the width direction so that the long side direction of the substrate FS wound on the rotating drum DR is always orthogonal to the center axis (rotation axis) AXo of the rotating drum DR, and appropriately adjust the parallelism of the rotation axis of the above-mentioned roller of the edge position controller EPC and the Y axis so as to correct the tilt error in the moving direction of the substrate FS.

The rotating drum DR has a central axis AXo extending in the Y direction and in a direction intersecting the direction of gravity, and a cylindrical outer peripheral surface with a certain radius from the central axis AXo. The rotating drum DR supports a portion of the substrate FS in the longitudinal direction in accordance with the outer peripheral surface (circumferential surface), and at the same time rotates around the central axis AXo to transport the substrate FS in the +X direction. The rotating drum DR supports the exposure area (portion) on the substrate FS for the light beam LB (point light SP) from the drawing head 16 with its circumferential surface. On both sides of the rotating drum DR in the Y direction, there is a shaft Sft supported by annular bearings in a manner that allows the rotating drum DR to rotate around the central axis AXo. The shaft Sft rotates around the central axis AXo by being given a torque from a rotational drive source (for example, composed of a motor and a reduction mechanism) not shown in the figure and controlled by the control device 18. In addition, for the sake of convenience, the plane containing the central axis AXo and parallel to the YZ plane is referred to as the central plane Poc.

The drive rollers R2 and R3 are arranged at predetermined intervals along the conveying direction (+X direction) of the substrate FS, and impart predetermined slack (play) to the exposed substrate FS. Like the drive roller R1, the drive rollers R2 and R3 rotate while holding the front and back surfaces of the substrate FS to convey the substrate FS toward the processing device PR2. The drive rollers R2 and R3 are arranged on the downstream side (+X direction side) of the conveying direction relative to the rotating drum DR, and the drive roller R2 is arranged on the upstream side (-X direction side) of the conveying direction relative to the drive roller R3. The tension adjustment rollers RT1 and RT2 are applied in the -Z direction to impart predetermined tension in the longitudinal direction to the substrate FS wound and supported on the rotating drum DR. In this way, the tension in the longitudinal direction imparted to the substrate FS hung on the rotating drum DR can be stabilized within a predetermined range. In addition, the control device 18 rotates the drive rollers R1 to R3 by controlling a rotation drive source (for example, composed of a motor and a reducer) not shown in the figure.

The light source device 14 has a light source (pulse light source) and emits a pulsed light beam (pulse light, laser) LB. The light beam LB is ultraviolet light having a peak wavelength in a band below 370nm, and the oscillation frequency (luminous frequency) of the light beam LB is Fs. The light beam LB emitted by the light source device 14 is incident on the drawing head 16. The light source device 14 emits and emits the light beam LB at a luminous frequency Fs in accordance with the control of the control device 18. The structure of the light source device 14 will be described in detail later, but it is composed of a semiconductor laser element that generates pulse light in the infrared band, an optical fiber amplifier, a wavelength conversion element (higher harmonic generation element) that converts the amplified pulse light in the infrared band into pulse light in the ultraviolet band, etc., and a fiber amplifier laser light source that obtains high-brightness ultraviolet pulse light with an oscillation frequency Fs of hundreds of MHz and a luminous time of one pulse of light of picoseconds can also be used.

The drawing head 16 has a plurality of scanning units Un (U1 to U6) for the light beam LB to be incident on respectively. The drawing head 16 draws a predetermined pattern on a portion of the substrate FS supported by the circumferential surface of the rotating drum DR of the substrate conveying mechanism 12 through the plurality of scanning units (drawing units) U1 to U6. The drawing head 16 is a so-called multi-beam type drawing head 16 in which a plurality of scanning units U1 to U6 having the same structure are arranged. The drawing head 16 repeatedly performs pattern exposure for electronic devices on the substrate FS, so that a plurality of exposure areas (electronic device formation areas) W of the exposed pattern are provided at predetermined intervals along the long side direction of the substrate FS (refer to FIG5 ). The control device 18 controls each part of the exposure device EX so that each part performs processing. The control device 18 includes a computer and a storage medium storing a program, and the computer functions as the control device 18 of the first embodiment by executing the program stored in the storage medium.

FIG2 illustrates a support frame (device column) 30 that supports the multiple scanning units (drawing units) Un and the rotating drum DR of the drawing head 16. The support frame 30 includes a main frame 32, a three-point support portion 34, and a drawing head support portion 36. The support frame 30 is housed within a temperature control chamber ECV. The main frame 32 rotatably supports the rotating drum DR and tension adjustment rollers RT1 (not shown) and RT2 via annular bearings. The three-point support portion 34 is provided at the upper end of the main frame 32 and supports the drawing head support portion 36, which is located above the rotating drum DR, at three points.

The drawing head support portion 36 is used to support the scanning units Un (U1 to U6) of the drawing head 16. The drawing head support portion 36 supports the scanning units U1, U3, and U5 on the downstream side (+X direction side) in the conveying direction relative to the central axis AXo of the rotating drum DR and arranged along the width direction of the substrate FS (refer to Figure 1). In addition, the drawing head support portion 36 supports the scanning units U2, U4, and U6 on the upstream side (-X direction side) in the conveying direction relative to the central axis AXo and arranged along the width direction (Y direction) of the substrate FS (refer to Figure 1). In addition, here, if the scanning width in the Y direction (scanning range of the spot light SP, drawing line SLn) achieved by one scanning unit Un is about 20 to 50 mm as an example, then by arranging a total of six scanning units Un, namely three odd-numbered scanning units U1, U3, and U5 and three even-numbered scanning units U2, U4, and U6, along the Y direction, the width of the drawable Y direction can be expanded to about 120 to 300 mm.

FIG3 illustrates the configuration of the drawing head 16. In the first embodiment, the exposure apparatus EX includes two light source devices 14 (14a, 14b). The drawing head 16 includes a plurality of scanning units U1 to U6, a light guide optical system (light beam switching component) 40a that guides the light beam LB from the light source device 14a to the plurality of scanning units U1, U3, and U5, and a light guide optical system (light beam switching component) 40b that guides the light beam LB from the light source device 14b to the plurality of scanning units U2, U4, and U6.

First, the light introducing optical system (light beam switching member) 40a will be described using Fig. 4. Since the light introducing optical systems 40a and 40b have the same structure, the light introducing optical system 40a will be described here and the description of the light introducing optical system 40b will be omitted.

The light-introducing optical system 40a includes, from the light source device 14 (14a) side, a focusing lens 42, a collimating lens 44, a reflector 46, a focusing lens 48, a selection optical element 50, a reflector 52, a collimating lens 54, a focusing lens 56, a selection optical element 58, a reflector 60, a collimating lens 62, a focusing lens 64, a selection optical element 66, a reflector 68 and an absorber 70.

The condenser lens 42 and the collimator lens 44 are used to amplify the light beam LB emitted from the light source device 14a. Specifically, the condenser lens 42 first converges the light beam LB to a focal position on the rear side of the condenser lens 42, and the collimator lens 44 converts the light beam LB, which is converged and then diverged by the condenser lens 42, into parallel light with a predetermined beam diameter (e.g., several mm).

The reflector 46 reflects the light beam LB, which has been parallelized by the collimating lens 44, and directs it toward the selection optical element 50. The condenser lens 48 converges the light beam LB incident on the selection optical element 50 so that it forms a beam waist within the selection optical element 50. The selection optical element 50 is transmissive to the light beam LB and, for example, employs an acousto-optic modulator (AOM). When an ultrasonic signal (high-frequency signal) is applied to the AOM, the incident light beam LB (zero-order light) is diffracted at a diffraction angle corresponding to the high-frequency frequency to generate first-order diffracted light as an outgoing light beam (light beam LBn). In the first embodiment, the light beams LBn emitted from the plurality of selection optical elements 50, 58, and 66 as primary diffracted light and incident on the corresponding scanning units U1, U3, and U5 are represented by LB1, LB3, and LB5, respectively. Each selection optical element 50, 58, and 66 is treated as an element that functions to deflect the optical path of the light beam LB from the light source device 14 (14a). The structure, function, and effects of each selection optical element 50, 58, and 66 can be identical. The selection optical elements 50, 58, and 66 turn on and off the generation of diffracted light that diffracts the incident light beam LB in accordance with the on/off switching of a drive signal (high-frequency signal) from the control device 18.

To explain in detail, when the drive signal (high-frequency signal) from the control device 18 is off, the selection optical element 50 directs the incident light beam LB toward the next-stage selection optical element 58. On the other hand, when the drive signal (high-frequency signal) from the control device 18 is on, the selection optical element 50 diffracts the incident light beam LB and directs the light beam LB1, which is the primary diffracted light, toward the reflector 52. The reflector 52 reflects the incident light beam LB1 and directs it toward the collimating lens 100 of the scanning unit U1. Specifically, by switching (driving) the selection optical element 50 on/off by the control device 18, the selection optical element 50 switches whether the light beam LB1 is incident on the scanning unit U1.

Between the selection optical element 50 and the selection optical element 58, there are provided in the following order: a collimating lens 54 for restoring the light beam LB irradiated to the selection optical element 58 into parallel light, and a focusing lens 56 for converging (shrinking) the light beam LB that has become parallel light after passing through the collimating lens 54 again in a manner that forms a beam waist within the selection optical element 58.

The selection optical element 58, like the selection optical element 50, is transmissive to the light beam LB and, for example, employs an acousto-optic modulator (AOM). When the drive signal (high-frequency signal) sent from the control device 18 is off, the selection optical element 58 transmits the incident light beam LB as it is, irradiating the selection optical element 66. When the drive signal (high-frequency signal) sent from the control device 18 is on, the selection optical element 58 diffracts the incident light beam LB and irradiates the light beam LB3, which is the primary diffracted light, toward the reflector 60. The reflector 60 reflects the incident light beam LB3 and irradiates it toward the collimating lens 100 of the scanning unit U3. Specifically, by switching the selection optical element 58 on and off by the control device 18, the selection optical element 58 switches whether the light beam LB3 is incident on the scanning unit U3.

Between the selection optical element 58 and the selection optical element 66, there are provided in the following order: a collimating lens 62 for restoring the light beam LB irradiated to the selection optical element 66 into parallel light, and a focusing lens 64 for converging (shrinking) the light beam LB that has become parallel light after passing through the collimating lens 62 again in a manner that forms a beam waist within the selection optical element 66.

Like the selection optical element 50, the selection optical element 66 is transmissive to the light beam LB and, for example, employs an acousto-optic modulator (AOM). When the drive signal (high-frequency signal) from the control device 18 is off, the selection optical element 66 directs the incident light beam LB toward the absorber 70. When the drive signal (high-frequency signal) from the control device 18 is on, the selection optical element 66 diffracts the incident light beam LB and directs the light beam LB5, which is the primary diffracted light, toward the reflector 68. The reflector 68 reflects the incident light beam LB5 and directs it toward the collimating lens 100 of the scanning unit U5. Specifically, by switching the selection optical element 66 on and off by the control device 18, the selection optical element 66 switches whether the light beam LB5 is incident on the scanning unit U5. The absorber 70 is a light trap that absorbs the light beam LB and prevents it from leaking to the outside.

The light introduction optical system 40b will be briefly described. The selection optical elements 50, 58, and 66 of the light introduction optical system 40b switch whether the light beam LB is incident on the scanning units U2, U4, and U6. In this case, the reflectors 52, 60, and 68 of the light introduction optical system 40b reflect the light beams LB2, LB4, and LB6 emitted from the selection optical elements 50, 58, and 66 and direct them toward the collimating lenses 100 of the scanning units U2, U4, and U6.

Furthermore, the efficiency of generating first-order diffracted light by an actual acousto-optic modulator (AOM) is approximately 80% of that of zero-order light. Therefore, the intensity of the light beams LB1 (LB2), LB3 (LB4), and LB5 (LB6) deflected by the selective optical elements 50, 58, and 66, respectively, is reduced compared to the original light beam LB. Furthermore, when any of the selective optical elements 50, 58, and 66 is in the on state, approximately 20% of the zero-order light remains, remaining undiffracted and traveling straight ahead. However, this light is ultimately absorbed by the absorber 70.

Next, the multiple scanning units Un (U1-U6) shown in Figure 3 will be described. Each scanning unit Un projects a light beam LBn from a light source device 14 (14a, 14b) onto the illuminated surface of the substrate FS, converging into a point light SP. Simultaneously, a rotating polygonal mirror PM causes this point light SP to be one-dimensionally scanned along a predetermined linear drawing line (scanning line) SLn on the illuminated surface of the substrate FS. The drawing line SLn of scanning unit U1 is denoted by SL1, and similarly, the drawing lines SLn of scanning units U2-U6 are denoted by SL2-SL6.

FIG5 is a diagram showing the drawing lines SLn (SL1 to SL6) that are scanned by the spot light SP through the scanning units Un (U1 to U6). As shown in FIG5 , the scanning area is shared by the scanning units Un (U1 to U6) in such a manner that all the scanning units of the plurality of scanning units Un (U1 to U6) cover the entire width direction of the exposure area W. Thus, each scanning unit Un (U1 to U6) can draw a pattern in each of the plurality of areas divided along the width direction of the substrate FS. The lengths of the drawing lines SLn (SL1 to SL6) are the same in principle. That is, the scanning distances of the spot light SP of the light beam LBn scanned along the drawing lines SL1 to SL6 are the same in principle. In addition, when it is desired to increase the width of the exposure area W, this can be addressed by increasing the length of the drawing line SLn itself or by increasing the number of scanning units Un provided in the Y direction.

Furthermore, each actual drawing line SLn (SL1 to SL6) is set to be slightly shorter than the maximum length that the spot light SP can actually scan on the irradiated surface. For example, assuming that the maximum length of the drawing line SLn that can draw a pattern is 30 mm when the drawing magnification in the main scanning direction (Y direction) is the initial value (without magnification correction), the drawing line SLn is provided with a margin of approximately 0.5 mm on both the scanning start point side and the scanning end point side, so that the maximum scanning length of the spot light SP on the irradiated surface is set to approximately 31 mm. By setting it in this way, the position of the 30 mm drawing line SLn in the main scanning direction or the drawing magnification can be fine-tuned within the maximum scanning length of 31 mm of the spot light SP. The maximum scanning length of the spot light SP is not limited to 31 mm and is mainly determined by the aperture of the fθ lens FT (see FIG. 3 ) located after the polygon mirror (rotating polygon mirror) PM in the scanning unit Un, and can also be greater than 31 mm.

Multiple drawing lines (scanning lines) SL1-SL6 are arranged in two rows in the circumferential direction of the rotating drum DR, sandwiching a center plane Poc. Drawing lines SL1, SL3, and SL5 are located on the substrate FS downstream (+X direction) in the conveyance direction relative to the center plane Poc. Drawing lines SL2, SL4, and SL6 are located on the substrate FS upstream (-X direction) in the conveyance direction relative to the center plane Poc. Each drawing line SLn (SL1-SL6) is approximately parallel to the width of the substrate FS, i.e., the central axis AXo of the rotating drum DR, and is shorter than the width of the substrate FS.

Drawing lines SL1, SL3, and SL5 are arranged at predetermined intervals along the width direction (scanning direction, Y direction) of the substrate FS. Drawing lines SL2, SL4, and SL6 are also arranged at predetermined intervals along the width direction (scanning direction, Y direction) of the substrate FS. At this time, drawing line SL2 is arranged between drawing line SL1 and drawing line SL3 in the width direction of the substrate FS. Similarly, drawing line SL3 is arranged between drawing line SL2 and drawing line SL4 in the width direction of the substrate FS. Drawing line SL4 is arranged between drawing line SL3 and drawing line SL5 in the width direction of the substrate FS. Drawing line SL5 is arranged between drawing line SL4 and drawing line SL6 in the width direction of the substrate FS. That is, drawing lines SL1 to SL6 are arranged to cover the entire width direction of the exposure area W to be drawn on the substrate FS.

The scanning direction of the spot light SP of the light beam LBn (LB1, LB3, LB5) scanning along the odd-numbered drawing lines SL1, SL3, and SL5 is one-dimensional and in the same direction. The scanning direction of the spot light SP of the light beam LBn (LB2, LB4, LB6) scanning along the even-numbered drawing lines SL2, SL4, and SL6 is one-dimensional and in the same direction. The scanning direction of the light beam LBn (spot light SP) scanning along the drawing lines SL1, SL3, and SL5 is opposite to the scanning direction of the light beam LBn (spot light SP) scanning along the drawing lines SL2, SL4, and SL6. Specifically, the scanning direction of the light beam LBn (spot light SP) scanning along the drawing lines SL2, SL4, and SL6 is the +Y direction, while the scanning direction of the light beam LBn (spot light SP) scanning along the drawing lines SL1, SL3, and SL5 is the -Y direction. This is due to the use of a polygonal mirror PM rotating in the same direction as the polygonal mirror PM of the scanning units U1 to U6. As a result, the drawing start position (the position of the drawing start point (scanning start point)) of the drawing lines SL1, SL3, and SL5 is adjacent to (or partially overlaps with) the drawing start position of the drawing lines SL2, SL4, and SL6 in the Y direction. In addition, the drawing end position (the position of the drawing end point (scanning end point)) of the drawing lines SL3 and SL5 is adjacent to (or partially overlaps with) the drawing end position of the drawing lines SL2 and SL4 in the Y direction. When each drawing line SLn is arranged so that the ends of the adjacent drawing lines SLn in the Y direction partially overlap, for example, it is sufficient as long as the length of each drawing line SLn, including the drawing start position or the drawing end position, overlaps in the Y direction by less than a few percent.

Furthermore, the width of the drawing line SLn in the sub-scanning direction is sized to correspond to the size (diameter) of the spotlight SP. For example, if the spotlight SP is 3 μm in size, the width of the drawing line SLn in the sub-scanning direction is also 3 μm. The spotlight SP can also be projected along the drawing line SLn so that it overlaps by a predetermined length (e.g., half the size of the spotlight SP). Furthermore, when adjacent drawing lines SLn in the Y direction (e.g., drawing lines SL1 and SL2) are adjacent to each other (contacting each other), the overlap only needs to be by a predetermined length (e.g., half the size of the spotlight SP).

In the case of the first embodiment, since the light beam LB from the light source device 14 is a pulsed light, the point light SP projected on the drawing line SLn during the main scanning period is discrete in accordance with the oscillation frequency Fs of the light beam LB. Therefore, it is necessary to make the point light SP projected by one pulse of the light beam LB and the point light SP projected by the next pulse overlap in the main scanning direction. The amount of overlap is set according to the size of the point light SP, the scanning speed Vs of the point light SP, and the oscillation frequency Fs of the light beam LB. However, when the intensity distribution of the point light SP is approximately Gaussian, it is sufficient to overlap by about 1/e ² (or 1/2) relative to the effective diameter size determined by 1/e 2 (or 1/2) of the peak intensity of the point light SP. Therefore, in the sub-scanning direction (the direction orthogonal to the drawing line SLn), it is also desirable to set the substrate FS to move a distance of approximately 1/2 of the effective size of the point light SP between one scan and the next scan along the drawing line SLn. In addition, regarding the setting of the exposure amount of the photosensitive functional layer on the substrate FS, although it can be achieved by adjusting the peak value of the light beam LB (pulsed light), when the exposure amount wants to be increased without increasing the intensity of the light beam LB, it is only necessary to increase the overlap amount of the point light SP in the main scanning direction or the sub-scanning direction to more than 1/2 of the effective size by reducing the scanning speed Vs of the point light SP in the main scanning direction, increasing the oscillation frequency Fs of the light beam LB, and reducing the transport speed in the sub-scanning direction of the substrate FS.

Next, the structure of the scanning unit Un shown in Figure 3 will be described. Since each scanning unit U1 to U6 has the same structure, only the scanning unit U1 will be described here. Scanning unit U1 includes, following the reflector 52 shown in Figure 4, a collimating lens 100, a reflector 102, a focusing lens 104, an image drawing optical element 106, a collimating lens 108, a reflector 110, a cylindrical lens CYa, a reflector 114, a polygonal mirror (light scanning unit, deflection unit) PM, an fθ lens FT, a cylindrical lens CYb, and a reflector 122. The collimating lenses 100, 108, the reflectors 102, 110, 114, 122, the focusing lens 104, the cylindrical lenses CYa, CYb, and the fθ lens FT constitute an optical lens system.

In Figure 3, the reflector 102 reflects the incident light beam LB1 from the collimating lens 100 in the -Z direction, causing it to enter the image drawing optical element 106, which serves as an image drawing light modulator. The condenser lens 104 converges the incident light beam LB1 (a parallel light beam) entering the image drawing optical element 106, forming a beam waist within the image drawing optical element 106. The image drawing optical element 106 is transmissive to the light beam LB1 and, for example, employs an acousto-optic modulator (AOM). When the drive signal (high-frequency signal) from the control device 18 is off, the image drawing optical element 106 directs the incident light beam LB1 toward a shielding plate or absorber (not shown). When the drive signal (high-frequency signal) from the control device 18 is on, the incident light beam LB1 is diffracted and the primary diffracted light (the image drawing light beam, i.e., the light beam LB1 intensity-modulated according to the pattern data) is directed toward the reflector 110. The shielding plate and absorber are used to prevent the light beam LB1 from leaking to the outside.

Between the reflector 110 and the optical element 106 for drawing, there is a collimating lens 108 that converts the incident light beam LB1 into parallel light. The reflector 110 reflects the incident light beam LB1 toward the reflector 114 in the -X direction, and the reflector 114 reflects the incident light beam LB1 toward the polygon mirror PM. The polygon mirror (rotating polygon mirror) PM reflects the incident light beam LB1 in the -X direction toward the fθ lens FT having an optical axis parallel to the X axis. The polygon mirror PM deflects (reflects) the incident light beam LB1 within a plane parallel to the XY plane in order to scan the point light SP of the light beam LB1 on the irradiated surface of the substrate FS. Specifically, the polygon mirror PM has a rotation axis AXp extending in the Z direction and a plurality of reflection surfaces RP (eight reflection surfaces RP in the first embodiment) formed around the rotation axis AXp. By rotating the polygonal mirror PM in a predetermined direction of rotation with the rotation axis AXp as the center, the reflection angle of the pulsed light beam LB1 irradiated on the reflection surface RP can be continuously changed. Thus, by deflecting the reflection direction of the light beam LB1 through a reflection surface RP, the point light SP of the light beam LB1 irradiated on the irradiated surface of the substrate FS can be scanned in the scanning direction (the width direction of the substrate FS, the Y direction). In other words, the polygonal mirror PM deflects the incident light beam LB1 so that the point light SP is scanned along the drawing line (scanning line) SL1 shown in Figure 5. In addition, the polygonal mirror PM is rotated at a certain speed by a rotation drive source (for example, composed of a motor and a reduction mechanism, etc.) not shown in the figure. The rotation drive source is controlled by the control device 18.

Because the point light SP of the light beam LB1 can be scanned along the drawing line SL1 by one reflection surface RP of the polygon mirror PM, the maximum number of drawing lines SL1 scanned by the point light SP on the illuminated surface of the substrate FS during one rotation of the polygon mirror PM is eight, the same number as the reflection surfaces RP. As described above, the effective length of the drawing line SL1 (e.g., 30 mm) is set to a length less than the maximum scanning length (e.g., 31 mm) that the point light SP can scan by the polygon mirror PM. In the initial setting (design), the center point of the drawing line SL1 is set at the center of the maximum scanning length.

Furthermore, as an example, if the effective length of the drawing line SL1 is set to 30 mm, and the spot lights SP having an effective size of 3 μm are irradiated onto the irradiated surface of the substrate FS along the drawing line SL1 while overlapping each other by 1.5 μm, the number of spot lights SP irradiated in one scan (the number of pulses of the light beam LB from the light source device 14) is 20,000 (30 mm/1.5 μm). Furthermore, if the scanning time of the spot lights SP along the drawing line SL1 is set to 200 μsec, the pulsed spot lights SP must be irradiated 20,000 times during this period. Therefore, the light emission frequency Fs of the light source device 14 becomes Fs ≥ 20,000 times/200 μsec = 100 MHz.

Returning to the description of the structure of the scanning unit U1, the cylindrical lens CYa disposed between the reflector 110 and the reflector 114 converges (converges) the light beam LB1 on the reflection surface RP of the polygon mirror PM in the Z direction (non-scanning direction) perpendicular to the scanning direction into an elongated elliptical shape (slit shape) extending in a direction parallel to the XY plane. This cylindrical lens CYa suppresses the influence of tilting the reflection surface RP relative to the Z direction (Z axis) (in the presence of a surface tilt error), thereby preventing the irradiation position of the point light beam LB1 irradiating the substrate FS from being shifted in the conveying direction (X direction) of the substrate FS.

The light beam LB1 reflected by the polygon mirror PM is irradiated onto the fθ lens FT including the focusing lens. The fθ lens FT having an optical axis extending in the X-axis direction is a telecentric scanning lens that projects the light beam LB1 reflected by the polygon mirror PM onto the reflector 122 in a plane parallel to the XY plane in a manner parallel to the X-axis. The incident angle θ of the light beam LB1 onto the fθ lens FT changes according to the rotation angle (θ/2) of the polygon mirror PM. The fθ lens FT projects the light beam LB1 onto the image height position on the irradiated surface of the substrate FS that is proportional to the incident angle θ. If the focal length is fo and the image height position is y, the fθ lens FT has the relationship y=fo·θ. Therefore, the light beam LB1 (point light SP) can be scanned accurately and uniformly in the Y direction by the fθ lens FT. When the incident angle onto the fθ lens FT is 0 degrees, the light beam LB1 incident on the fθ lens FT travels along the optical axis of the fθ lens FT.

The light beam LB1 irradiated from the fθ lens FT is irradiated on the substrate FS as a point light SP via the reflector 122. The cylindrical lens CYb provided between the fθ lens FT and the reflector 122 makes the point light SP of the light beam LB1 converged on the substrate FS into a tiny circle with a diameter of several μm (for example, 3 μm), whose generatrix is parallel to the Y direction. Thus, a drawing line SL1 extending in the Y direction based on the point light (scanning point) SP is defined on the substrate FS (refer to FIG5 ). In the absence of the cylindrical lens CYb, the point light SP converged on the substrate FS becomes an elongated ellipse extending in a direction (X direction) orthogonal to the scanning direction (Y direction) due to the action of the cylindrical lens CYa in front of the polygonal mirror PM.

In this way, when the substrate FS is transported along the X direction, the point light SP of the light beam LB is scanned in the scanning direction (Y direction) by each scanning unit U1 to U6, thereby drawing a predetermined pattern on the substrate FS. The scanning units U1 to U6 are arranged on the drawing head support part 36 in a manner that scans different areas on the substrate FS. In addition, when the size of the point light SP on the substrate FS in the scanning direction (the length of the drawing line) is set to Ds and the scanning speed of the point light SP on the substrate FS (the relative scanning speed) is set to Vs, the oscillation frequency Fs of the light beam LB needs to satisfy the relationship Fs≥Vs/Ds. This is because the light beam LB is a pulsed light, so if the oscillation frequency Fs does not satisfy the relationship Fs≥Vs/Ds, the point light SP of the light beam LB will be irradiated onto the substrate FS at predetermined intervals (gaps). If the oscillation frequency Fs satisfies the relationship Fs ≥ Vs / Ds, the point light SP can be irradiated onto the substrate FS in a manner that overlaps with each other in the scanning direction. Therefore, even with a pulsed oscillating light beam LB, a substantially continuous straight line pattern along the scanning direction can be well drawn on the substrate FS. In addition, the faster the rotation speed of the polygon mirror PM becomes, the faster the scanning speed Vs of the point light SP becomes.

Figure 6 shows the relationship between the polygon mirror PM of each scanning unit U1 to U6 and the scanning direction of the multiple drawing lines SLn (SL1 to SL6). The reflector 114, polygon mirror PM, and fθ lens FT in the multiple scanning units U1, U3, and U5 and the multiple scanning units U2, U4, and U6 are symmetrical about the center plane Poc. Therefore, by rotating the polygon mirror PM of each scanning unit U1 to U6 in the same direction (to the left), each scanning unit U1, U3, and U5 causes the point light SP of the light beam LB to scan in the -Y direction from the drawing start position to the drawing end position, while each scanning unit U2, U4, and U6 causes the point light SP of the light beam LB to scan in the +Y direction from the drawing start position to the drawing end position. In addition, the scanning directions of the point light SP of the light beam LB of each scanning unit U1 to U6 can be adjusted to the same direction (+Y direction or -Y direction) by making the rotation direction of the polygonal mirror PM of each scanning unit U2, U4, and U6 opposite to the rotation direction of the polygonal mirror PM of each scanning unit U1, U3, and U5.

Here, since the polygon mirror PM rotates, the angle of the reflection surface RP also changes over time. Therefore, there is a limit to the rotation angle α of the polygon mirror PM at which the light beam LB incident on a specific reflection surface RP of the polygon mirror PM can be incident on the fθ lens FT.

Figure 7 is a diagram for explaining the rotation angle α of the polygonal mirror PM of the scanning unit Un, which is capable of deflecting (reflecting) the light beam LBn so that it is incident on the fθ lens FT. The rotation angle α is the maximum scanning rotation angle range of the polygonal mirror PM of the scanning unit Un, which is capable of scanning the point light SP on the irradiated surface of the substrate FS through one reflection surface RP. Hereinafter, the rotation angle α will be referred to as the maximum scanning rotation angle range. The period during which the polygonal mirror PM rotates within the maximum scanning rotation angle range α is the effective scanning period (maximum scanning time) of the point light SP. The maximum scanning rotation angle range α corresponds to the maximum scanning length of the above-mentioned drawing line SLn, and the larger the maximum scanning rotation angle range α becomes, the longer the maximum scanning length becomes. The rotation angle β represents the rotation angle from the angle of the polygonal mirror PM when the light beam LB starts to be incident on a specific reflection surface RP to the angle of the polygonal mirror PM when the incidence on the above-mentioned specific reflection surface RP ends. That is, the rotation angle β is the angle at which the polygonal mirror PM rotates by the amount of one surface of the reflection surface RP. The rotation angle β is determined by the number Np of the reflection surfaces RP of the polygonal mirror PM and can be expressed as β≈360/Np. Therefore, the above-mentioned specific reflection surface RP of the polygonal mirror PM of the scanning unit Un cannot cause the point light SP to scan the irradiated surface of the substrate FS, that is, the reflected light reflected by the above-mentioned specific reflection surface RP of the polygonal mirror PM cannot be incident on the fθ lens FT, and the non-scanning rotation angle range γ of the polygonal mirror PM is expressed by the relationship γ=β-α. The period during which the polygonal mirror PM rotates within the non-scanning rotation angle range γ becomes an invalid scanning period of the point light SP. In this non-scanning rotation angle range γ, the scanning unit Un cannot irradiate the light beam LBn onto the substrate FS. The rotation angle α and the non-scanning rotation angle range γ have a relationship as shown in formula (1).

γ＝(360 degrees/Np)－α…(1)

(where N is the number of reflecting surfaces RP of the polygon mirror PM)

In the first embodiment, since the polygon mirror PM has eight reflecting surfaces RP, N = 8. Therefore, equation (1) can be expressed by equation (2).

γ＝45 degrees－α…(2)

The maximum scanning rotation angle range α varies depending on conditions such as the distance between the polygon mirror PM and the fθ lens FT. For example, if the maximum scanning rotation angle range α is set to 15 degrees, the non-scanning rotation angle range γ is 30 degrees, and the scanning efficiency of the polygon mirror PM is α/β = 1/3 in Figure 7. In other words, while the polygon mirror PM of the scanning unit Un rotates within the non-scanning rotation angle range γ (30 degrees), the light beam LBn incident on the polygon mirror PM is useless.

Therefore, in the first embodiment, the scanning unit Un into which the light beam LB from one light source device 14 is incident is switched, and the light beam LB is periodically distributed to the three scanning units Un, thereby improving scanning efficiency. In other words, by staggering the drawing periods (scanning periods during which the spot light SP is scanned) of the three scanning units Un, the light beam LB from the light source device 14 is not wasted, thereby improving scanning efficiency.

Furthermore, while the effective scanning period (effective drawing period), i.e., the maximum scanning rotation angle range α, is the range within which the light beam LBn can be incident on the fθ lens FT and the spot light SP can be effectively scanned on the drawing line SLn, the maximum scanning rotation angle range α also varies depending on, for example, the focal length of the front side of the fθ lens FT. In the case of the same eight-faceted polygonal mirror PM as described above, when the maximum scanning rotation angle range α is 10 degrees, according to equation (2), the non-drawing period, i.e., the non-scanning rotation angle range γ, becomes 35 degrees. In this case, the drawing scanning efficiency is approximately 1/4 (10/45). Conversely, when the maximum scanning rotation angle range α is 20 degrees, according to equation (2), the non-drawing period, i.e., the non-scanning rotation angle range γ, becomes 25 degrees. In this case, the drawing scanning efficiency is approximately 1/2 (20/45). Furthermore, when the scanning efficiency is 1/2 or greater, the number of scanning units Un that distribute the light beam LB may be two. In other words, the number of scanning units Un that can distribute the light beam LB is limited by the scanning efficiency.

FIG8 is a diagram schematically illustrating the optical path between the light-introducing optical system 40a and the plurality of scanning units U1, U3, and U5. When the drive signal (high-frequency signal) applied from the control device 18 to the selection optical element (AOM) 50 is on and the drive signal applied to the selection optical elements 58 and 66 is off, the selection optical element 50 diffracts the incident light beam LB. Consequently, the primary diffracted light diffracted by the selection optical element 50, i.e., the light beam LB1, is incident on the scanning unit U1 via the reflector 52, and the light beam LB does not enter the scanning units U3 and U5. Similarly, when the drive signal applied from the control device 18 to the selection optical element (AOM) 58 is on and the drive signal applied to the selection optical elements 50 and 66 is off, the light beam LB that has passed through the off-state selection optical element 50 enters the selection optical element 58, and the selection optical element 58 diffracts the incident light beam LB. As a result, the first-order diffracted light beam LB3 diffracted by the selection optical element 58 enters the scanning unit U3 via the reflector 60, and the light beam LB does not enter the scanning units U1 and U5. Furthermore, when the drive signal applied from the control device 18 to the selection optical element (AOM) 66 is on and the drive signal applied to the selection optical elements 50 and 58 is off, the light beam LB that has passed through the off-state selection optical elements 50 and 58 enters the selection optical element 66, which diffracts the incident light beam LB. As a result, the first-order diffracted light beam LB5 diffracted by the selection optical element 66 enters the scanning unit U5 via the reflector 68, and the light beam LB does not enter the scanning units U1 and U3.

By arranging the plurality of selection optical elements 50, 58, and 66 for guiding the light into the optical system 40a in series along the direction of travel of the light beam LB from the light source device 14a in this manner, the plurality of selection optical elements 50, 58, and 66 can switchably select whether or not the light beam LBn (LB1, LB3, and LB5) is incident on a particular scanning unit Un among the plurality of scanning units U1, U3, and U5. The control device 18 controls the plurality of selection optical elements 50, 58, and 66 so that the scanning unit Un on which the light beam LB is incident is periodically switched, for example, in the order of scanning unit U1, scanning unit U3, scanning unit U5, and scanning unit U1. In other words, the plurality of selection optical elements 50, 58, and 66 are switched so that the light beam LBn (LB1, LB3, and LB5) is incident on each of the plurality of scanning units U1, U3, and U5 in sequence for a predetermined scanning period.

During the period when the light beam LB1 is incident on the scanning unit U1, the rotation of the polygonal mirror PM of the scanning unit U1 is controlled by the control device 18 so that the incident light beam LB1 can be reflected toward the fθ lens FT. That is, the period during which the light beam LB1 is incident on the scanning unit U1 is synchronized with the period during which the point light SP of the light beam LB1 is scanned by the scanning unit U1 (the maximum scanning rotation angle range α in FIG7 ). In other words, the polygonal mirror PM of the scanning unit U1 deflects the light beam LB1 in synchronization with the period during which the light beam LB1 is incident so that the point light SP of the light beam LB1 incident on the scanning unit U1 is scanned along the drawing line SL1. Similarly, the polygonal mirror PM of the scanning units U3 and U5 is also controlled by the control device 18 so that the incident light beams LB3 and LB5 can be reflected toward the fθ lens FT during the period during which the light beams LB1 are incident on the scanning units U3 and U5. That is, the period during which the light beams LB3 and LB5 are incident on the scanning units U3 and U5 is synchronized with the period during which the point light SP of the light beams LB3 and LB5 is scanned by the scanning units U3 and U5. In other words, the polygon mirror PM of the scanning units U3 and U5 deflects the light beams LB3 and LB5 in synchronization with the period during which the light beams LB3 and LB5 are incident, so that the point light SP of the light beam LB incident on the scanning units U3 and U5 is scanned along the drawing lines SL3 and SL5.

In this manner, the light beam LB from one light source device 14a is supplied to one scanning unit Un among the three scanning units U1, U3, and U5 in a time-sharing manner. Therefore, the polygon mirrors PM of each scanning unit U1, U3, and U5 are controlled and driven to rotate so that the rotation speeds of the polygon mirrors PM are consistent and their rotational angular positions maintain a certain angular difference (maintain a phase difference). A specific example of this control will be described later.

Furthermore, the control device 18 controls the on/off switching of the drive signal (high-frequency signal) supplied to the drawing optical element 106 of each scanning unit U1, U3, and U5 based on pattern data (drawing data). This pattern data specifies the pattern drawn on the substrate FS by the point light SP of the light beam LB1, LB3, and LB5 emitted from each scanning unit U1, U3, and U5. Consequently, the drawing optical element 106 of each scanning unit U1, U3, and U5 can diffract the incident light beam LB1, LB3, and LB5 based on the on/off drive signal, thereby modulating the intensity of the point light SP. This pattern data is generated as bitmap data, for example, by setting one dot (pixel) of the drawing pattern to 3×3 μm. For each dot, the binary data is set to "1" when the drive signal is on (drawing) and "0" when the drive signal is off (not drawing). This data is temporarily stored in the memory (RAM) for each scanning unit Un.

The pattern data set for each scanning unit Un is further described in detail. The pattern data (drawing data) is bitmap data composed of data of a plurality of pixels (hereinafter referred to as pixel data) that are two-dimensionally decomposed with the direction along the scanning direction of the point light SP (main scanning direction, Y direction) as the row direction and the direction along the conveying direction of the substrate FS (sub-scanning direction, X direction) as the column direction. The pixel data is 1-bit data of "0" or "1". The pixel data of "0" indicates that the intensity of the point light SP irradiated onto the substrate FS is at a low level (low level), and the pixel data of "1" indicates that the intensity of the point light SP irradiated onto the substrate FS is at a high level (high level). One column of the pattern data corresponds to one drawing line SLn (SL1 to SL6), and the intensity of the point light SP projected onto the substrate FS along one drawing line SLn (SL1 to SL6) is modulated according to one column of the pixel data. The pixel data of this column is called serial data (serial data) (drawing information) DLn. In other words, the pattern data is bitmap data in which serial data DLn are arranged in the column direction. Serial data DLn of the pattern data of scanning unit U1 is sometimes represented by DL1, and similarly, serial data DLn of the pattern data of scanning units U2 to U6 are sometimes represented by DL2 to DL6.

Based on the pattern data (serial data DLn consisting of "0" and "1") of the scanning unit Un on which the light beam LBn is incident, the control device 18 inputs an On/Off drive signal to the drawing optical element (AOM) 106 of the scanning unit Un on which the light beam LBn is incident. If the driving signal "On" is input to the drawing optical element 106, the incident light beam LBn is diffracted and irradiated toward the reflector 110. If the driving signal "Off" is input to the drawing optical element 106, the incident light beam LBn is irradiated toward the above-mentioned shielding plate or the above-mentioned absorber (not shown). As a result, with respect to the scanning unit Un on which the light beam LBn is incident, if the driving signal "On" is input to the drawing optical element 106, the point light SP of the light beam LBn is irradiated onto the substrate FS (the intensity of the point light SP becomes high). If the driving signal "Off" is input to the drawing optical element 106, the point light of the light beam LBn is not irradiated onto the substrate FS (the intensity of the point light SP becomes 0). Therefore, the scanning unit Un to which the light beam LBn is incident can draw a pattern based on the pattern data on the substrate FS along the drawing line SLn.

For example, when light beam LB3 is incident on scanning unit U3, control device 18 switches (drives) scanning unit U3's drawing optical element 106 on and off based on the pattern data of scanning unit U3. This enables scanning unit U3 to draw a pattern based on the pattern data along drawing line SL3 on substrate FS. In this manner, each scanning unit U1, U3, and U5 can modulate the intensity of spot light (scanning point) SP along drawing lines SL1, SL3, and SL5, thereby drawing a pattern based on the pattern data on substrate FS.

While FIG8 is used to illustrate the operation of the light introduction optical system 40a and the plurality of scanning units U1, U3, and U5, the same applies to the light introduction optical system 40b and the plurality of scanning units U2, U4, and U6. To briefly explain, the control device 18 controls the plurality of selection optical elements 50, 58, and 66 so that the even-numbered scanning units Un, on which the light beam LBn from the light source device 14b is incident, are switched in the order, for example, scanning unit U2 → scanning unit U4 → scanning unit U6 → scanning unit U2. In other words, the switching is performed so that the light beam LBn is sequentially incident on the plurality of scanning units U2, U4, and U6 for a predetermined scanning period. Under the control of the control device 18, the polygon mirror PM of each scanning unit U2, U4, and U6 deflects the light beam LBn in synchronization with the incident period, so that the spot light SP of the incident light beam LBn scans along the drawing lines SL2, SL4, and SL6. In addition, the control device 18 controls the drawing optical element (AOM) 106 of the scanning unit Un (U2, U4, U6) based on the pattern data (serial data DLn (DL2, DL4, DL6) composed of "0" and "1") of the scanning unit Un (U2, U4, U6) on which the light beam LBn (LB2, LB4, LB6) is incident, so that each scanning unit U2, U4, U6 can draw a pattern based on the pattern data along the drawing lines SL2, SL4, SL6 on the substrate FS.

As described above, in the first embodiment, since a plurality of selection optical elements 50, 58, and 66 are arranged in series along the traveling direction of the light beam LB from the light source device 14a (14b), the light beam LBn can be selectively incident on one of the scanning units Un among the plurality of scanning units U1, U3, and U5 (scanning units U2, U4, and U6) in a time-sharing manner through the plurality of selection optical elements 50, 58, and 66, and the light beam LB will not become useless, thereby improving the utilization efficiency of the light beam LB.

In addition, the rotation speed and rotation phase of the polygonal mirror PM of each of the multiple (here three) scanning units Un are synchronized with each other, and, in synchronization with the period during which the light beam LBn is incident on each scanning unit Un through the multiple selection optical elements 50, 58, and 66, the polygonal mirror PM deflects the light beam LBn in such a manner that the point light SP is scanned on the substrate FS. Therefore, the light beam LB will not become useless, and the scanning efficiency can be improved.

Furthermore, the optical elements (AOM) 50, 58, and 66 can be turned on only during a single scan of the point light SP by the respective polygonal mirrors PM of the scanning units Un. For example, if the number of reflection surfaces of the polygonal mirror PM is Np and the rotational speed Vp of the polygonal mirror PM is rpm, the time Tss corresponding to the rotation angle β of one side of the reflection surface RP of the polygonal mirror PM is Tss = 60/(Np·Vp) (seconds). For example, when the number of reflection surfaces Np is 8 and the rotational speed Vp is 30,000, one rotation of the polygonal mirror PM takes 2 milliseconds, and the time Tss is 0.25 milliseconds. Converting this to a frequency of 4 kHz, this means that compared to the acousto-optic modulator (optical element 106 for drawing) used to modulate the ultraviolet wavelength light beam LB at a high speed of several tens of MHz in response to pattern data, it can have a significantly lower response frequency. Therefore, the selection optical elements (AOMs) 50, 58, and 66 can use those with a large diffraction angle for the first-order diffracted light beams LBn (LB1 to LB6) deflected from the incident light beam LB (zero-order light). This facilitates the arrangement of the reflective mirrors 52, 60, and 68 (see Figures 3 and 4) that guide the light beams LBn (LB1 to LB6), which have been deflected from the path of the light beam LB that has linearly passed through the selection optical elements 50, 58, and 66, toward the scanning unit Un.

[Modification of the First Embodiment]

The first embodiment described above can be modified as follows: In the first embodiment described above, the light beam LB is distributed to three scanning units Un, but in this modification, the light beam LB from one light source device 14 is distributed to five scanning units Un.

FIG9 illustrates the structure of the drawing head 16 in a modified example of the first embodiment. In this modified example, a single light source device 14 is used, and the drawing head 16 includes five scanning units Un (U1 to U5). Components identical to those in the first embodiment are denoted by the same reference numerals or omitted from illustration, and only the differences are described. FIG9 also omits the cylindrical lens CYb shown in FIG3 .

In this modified example, instead of the light introducing optical systems 40a and 40b, a light introducing optical system (light beam switching component) 130 is used. As shown in FIG10 , the light introducing optical system 130 includes, in addition to the condensing lens 42, collimating lens 44, reflecting mirror 46, condensing lens 48, selection optical element 50, reflecting mirror 52, collimating lens 54, condensing lens 56, selection optical element 58, reflecting mirror 60, collimating lens 62, condensing lens 64, selection optical element 66, reflecting mirror 68, and absorber 70 shown in FIG4 , a selection optical element 132, reflecting mirror 134, collimating lens 136, condensing lens 138, selection optical element 140, reflecting mirror 142, collimating lens 144, and condensing lens 146.

The selection optical element 132, the collimating lens 136, and the condensing lens 138 are provided in this order between the condensing lens 56 and the selection optical element 58. Therefore, in this modified example, when the drive signal (high-frequency signal) from the control device 18 is off, the selection optical element 50 transmits the incident light beam LB as it is and irradiates the light beam toward the selection optical element 132. The condensing lens 56 condenses the light beam LB incident on the selection optical element 132 so that it forms a beam waist within the selection optical element 132.

The selection optical element 132 is transmissive to the light beam LB and, for example, is an acousto-optic modulator (AOM). When the drive signal from the control device 18 is off, the selection optical element 132 transmits the incident light beam LB directly and irradiates the light beam LB toward the selection optical element 58. When the drive signal (high-frequency signal) from the control device 18 is on, the selection optical element 132 irradiates the light beam LB2, a primary diffracted light beam from the incident light beam LB, toward the reflector 134. The reflector 134 reflects the incident light beam LB2, causing it to enter the collimating lens 100 of the scanning unit U2. In other words, by switching the selection optical element 132 on and off by the control device 18, the selection optical element 132 switches whether the light beam LB2 is incident on the scanning unit U2. The collimator lens 136 collimates the light beam LB directed toward the selection optical element 58 , and the condenser lens 138 converges the light beam LB collimated by the collimator lens 136 to form a beam waist within the selection optical element 58 .

The selection optical element 140, the collimating lens 144, and the condensing lens 146 are provided in this order between the condensing lens 64 and the selection optical element 66. Therefore, in this modified example, when the drive signal from the control device 18 is off, the selection optical element 58 transmits the incident light beam LB as it is and irradiates the light beam toward the selection optical element 140. The condensing lens 64 condenses the light beam LB incident on the selection optical element 140 so that it forms a beam waist within the selection optical element 140.

The selection optical element 140 is transmissive to the light beam LB and is, for example, an acousto-optic modulator (AOM). When the drive signal from the control device 18 is off, the selection optical element 140 directs the incident light beam LB toward the selection optical element 66. When the drive signal (high-frequency signal) from the control device 18 is on, the selection optical element 140 directs the light beam LB4, a primary diffracted light beam from the incident light beam LB, toward the reflector 142. The reflector 142 reflects the incident light beam LB4 and directs it toward the collimating lens 100 of the scanning unit U4. Specifically, by switching the selection optical element 140 on and off by the control device 18, the selection optical element 140 switches whether the light beam LB4 is incident on the scanning unit U4. The collimator lens 144 collimates the light beam LB directed toward the selection optical element 66 , and the condenser lens 146 converges the light beam LB collimated by the collimator lens 144 to form a beam waist within the selection optical element 66 .

By arranging these multiple selection optical elements (AOMs) 50, 58, 66, 132, and 140 in series (inline), the light beam LBn can be directed toward any one scanning unit Un among the multiple scanning units U1 to U5. The control device 18 controls the multiple selection optical elements 50, 132, 58, 140, and 66 so that the scanning unit Un on which the light beam LBn is incident is periodically switched, for example, in the order of scanning unit U1 → scanning unit U2 → scanning unit U3 → scanning unit U4 → scanning unit U5 → scanning unit U1. In other words, the light beam LBn is sequentially directed toward each of the multiple scanning units U1 to U5 for a predetermined scanning period. Furthermore, under the control of the control device 18, the polygon mirror PM of each scanning unit U1 to U5 deflects the light beam LBn in synchronization with the incident period, so that the spot light SP of the incident light beam LBn scans along the trace lines SL1 to SL5. In addition, the control device 18 controls the drawing optical element (AOM) 106 of the scanning unit Un based on the pattern data (serial data DLn composed of "0" and "1") of the scanning unit Un on which the light beam LBn is incident, so that each scanning unit Un can draw a pattern based on the pattern data along the drawing line SLn on the substrate FS.

That is, in the case of this modification, each polygon mirror PM of the five scanning units U1 to U5 rotates synchronously in such a manner that the rotation angle position is phase-shifted by a certain angle each time. In addition, in the case of this modification, the light beam (laser) LB is allocated to the five scanning units U1 to U5 in a time-sharing manner. Therefore, the front focal length of the fθ lens FT and/or the number of reflection surfaces Np of the polygon mirror PM are set so that the angle range (rotation angle β in Figure 7) within which the light beam LBn can illuminate one reflection surface RP of the polygon mirror PM and the maximum deflection angle (angle 2α in Figure 7) of the light beam LBn reflected by the reflection surface RP and incident on the fθ lens FT satisfy β≥5α.

As described above, in this modification, the light beam LB will not be rendered useless, and the utilization efficiency of the light beam LB from the light source device 14 can be improved, thereby achieving an improvement in scanning efficiency. In addition, in this modification, the light beam LB from one light source device 14 is distributed to five scanning units Un, but the light beam LB from one light source device 14 can also be distributed to two scanning units Un, or it can be distributed to four or six or more scanning units Un. In this case, if the number of distributed scanning units Un is set to n, the front focal length of the fθ lens FT and/or the number Np of reflecting surfaces of the polygonal mirror PM are set in such a way that the angle range (rotation angle β in FIG. 7 ) in which the light beam LBn can illuminate one reflecting surface RP of the polygonal mirror PM and the maximum deflection angle (angle 2α in FIG. 7 ) of the light beam LB reflected by the reflecting surface RP and incident on the fθ lens FT satisfy β ≥ n × α. Furthermore, as described in the first embodiment, when the light beams LB from the two light source devices 14 (14a, 14b) are distributed to a plurality of scanning units Un, the distribution is not limited to three scanning units Un, and the light beams LB can be distributed to any number of scanning units Un. For example, the light beam LB from the light source device 14a can be distributed to five scanning units Un, and the light beam LB from the light source device 14b can be distributed to four scanning units Un.

[Second embodiment]

In the first embodiment described above, since a drawing optical element (AOM) 106 is provided in front of the polygonal mirror PM in each scanning unit Un, the number of drawing optical elements 106 used increases, and the cost increases. Therefore, in this second embodiment, a drawing light modulator (AOM) is provided on the optical path of the light beam LB from a light source device 14, and this single drawing light modulator is used to modulate the intensity of the light beam LBn irradiated from the multiple scanning units Un to the substrate FS to draw a pattern. That is, in the second embodiment, only one drawing light modulator (AOM) requiring high responsiveness is arranged in front of the multiple scanning units Un, and a selection optical element (AOM) whose responsiveness can be low is arranged on the side of each scanning unit Un.

FIG11 illustrates the structure of the drawing head 16 according to the second embodiment, and FIG12 illustrates the light introduction optical system 40a shown in FIG11 . Components identical to those in the first embodiment are designated by the same reference numerals, and only the differences will be described. FIG11 omits the cylindrical lens CYb shown in FIG3 . Since the light introduction optical systems 40a and 40b have the same structure, the light introduction optical system 40a will be described here, while the description of the light introduction optical system 40b will be omitted. As shown in FIG12 , the light introduction optical system 40a includes, in addition to the condenser lens 42, collimator lens 44, reflector 46, condenser lens 48, selection optical element 50, reflector 52, collimator lens 54, condenser lens 56, selection optical element 58, reflector 60, collimator lens 62, condenser lens 64, selection optical element 66, reflector 68, and absorber 70 shown in FIG4 , an image drawing optical element (AOM) 150 serving as an image drawing light modulator, a collimator lens 152, condenser lens 154, and an absorber 156. In the second embodiment, as shown in FIG11 , each of the scanning units U1 to U6 does not include the image drawing optical element 106 as in the first embodiment.

The image drawing optical element 150, the collimating lens 152, and the condensing lens 154 are provided in this order between the condensing lens 48 and the selection optical element 50. Therefore, in the second embodiment, the reflector 46 reflects the light beam LB, which has been parallelized by the collimating lens 44, toward the image drawing optical element 150. The condensing lens 48 converges the light beam LB incident on the image drawing optical element 150 so that it forms a beam waist within the image drawing optical element 150.

The drawing optical element 150 is transmissive to the light beam LB and, for example, is an acousto-optic modulator (AOM). The drawing optical element 150 is located closer to the light source device 14 (14a) than the primary selection optical element 50, which is located closest to the light source device 14 (14a) among the selection optical elements 50, 58, and 66. When the drive signal (high-frequency signal) from the control device 18 is off, the drawing optical element 150 directs the incident light beam LB toward the absorber 156. When the drive signal (high-frequency signal) from the control device 18 is on, the drawing optical element 150 directs the light beam LB diffracted from the incident light beam LB as primary diffracted light (drawing light beam) LB toward the primary selection optical element 50. The collimating lens 152 parallelizes the light beam LB directed to the selection optical element 50, and the focusing lens 154 converges the light beam LB, which has been parallelized by the collimating lens 152, to form a beam waist within the selection optical element 50.

As shown in FIG11 , the scanning units U1 to U6 include a collimating lens 100, a reflecting mirror 102, a reflecting mirror 110, a cylindrical lens CYa, a reflecting mirror 114, a polygon mirror PM, an fθ lens FT, a cylindrical lens CYb (not shown in FIG11 ), and a reflecting mirror 122. Furthermore, the scanning units U1 to U6 include a first shaping lens 158a and a second shaping lens 158b as beam shaping lenses. Specifically, in the second embodiment, the first shaping lens 158a and the second shaping lens 158b are provided in place of the condenser lens 104 and the collimating lens 108 of the first embodiment.

Figure 13 is a diagram schematically illustrating the optical path between the light-introducing optical system 40a and the multiple scanning units U1, U3, and U5 in Figure 12 . The control device 18 outputs an on/off drive signal (a high-frequency signal) to the image drawing optical element 150 of the light-introducing optical system 40a based on pattern data (serial data DL1, DL3, and DL6 consisting of "0" and "1"). This pattern data specifies the pattern drawn on the substrate FS by the spot light SP of the light beams LB1, LB3, and LB5 emitted from the scanning units U1, U3, and U5. Consequently, the image drawing optical element 150 of the light-introducing optical system 40a can diffract the incident light beam LB based on the on/off drive signal to modulate the intensity of the spot light SP (turning it on and off).

To explain in detail, the control device 18 inputs an On/Off drive signal to the drawing optical element 150 based on the pattern data of the scanning unit Un on which the light beam LBn is incident. If the drawing optical element 150 is input with an On drive signal (high-frequency signal), the incident light beam LB is diffracted and irradiated toward the selection optical element 50 (the intensity of the light beam LB incident on the selection optical element 50 becomes higher). On the other hand, if the drawing optical element 150 is input with an Off drive signal (high-frequency signal), the incident light beam LB is irradiated toward the absorber 156 (Figure 12) (the intensity of the light beam LB incident on the selection optical element 50 becomes 0). Therefore, the scanning unit Un on which the light beam LBn is incident can irradiate the intensity-modulated light beam LB toward the substrate FS along the drawing line SLn, and can draw a pattern based on the pattern data on the substrate FS.

For example, when the light beam LB3 is incident on the scanning unit U3, the control device 18 switches the drawing optical element 150 of the light-introducing optical system 40a on/off based on the pattern data of the scanning unit U3. As a result, the scanning unit U3 can irradiate the substrate FS with the intensity-modulated light beam LB along the drawing line SL3, and can draw a pattern based on the pattern data on the substrate FS. The scanning unit Un on which the light beam LBn is incident is switched sequentially, for example, in the order of scanning unit U1 → scanning unit U3 → scanning unit U5 → scanning unit U1. Therefore, the control device 18 similarly switches the pattern data that determines the on/off signal sent to the drawing optical element 150 of the light-introducing optical system 40a in the order of pattern data of scanning unit U1 → pattern data of scanning unit U3 → pattern data of scanning unit U5 → pattern data of scanning unit U1. Then, the control device 18 controls the drawing optical element 150 of the light-introducing optical system 40a based on the pattern data obtained by sequential switching. Thus, each scanning unit U1 , U3 , U5 can draw a pattern corresponding to the pattern data on the substrate FS by irradiating the substrate FS with the intensity-modulated light beam LB along the drawing lines SL1 , SL3 , SL5 .

The above describes in detail a portion of the structure and operation of a control system applicable to the second embodiment with reference to Figures 14 to 16 . Furthermore, the structure and operation described below are also applicable to the first embodiment. Figure 14 is a block diagram of an example rotation control system for the polygon mirror PM provided within each of the three scanning units U1, U3, and U5 shown in Figures 11 and 13 . Since the scanning units U1, U3, and U5 have identical structures, identical components are designated by the same reference numerals. Each scanning unit U1, U3, and U5 is equipped with an origin sensor OP1, OP3, or OP5 that photoelectrically detects the start timing of scanning the drawing lines (scanning lines) SL1, SL3, and SL5 generated on the substrate FS by the polygon mirror PM. The origin sensors OP1, OP3, and OP5 are photodetectors that project light onto the reflective surface RP of the polygon mirror PM and receive the reflected light. Whenever the spot light SP reaches a position immediately before the scanning start point of the drawing lines SL1, SL3, and SL5, they output pulsed origin signals SZ1, SZ3, and SZ5, respectively.

The timing measurement unit 180 receives the origin signals SZ1, SZ3, and SZ5 as input, and measures whether the generation timing of each of the origin signals SZ1, SZ3, and SZ5 is within a predetermined tolerance range (time interval). If an error deviates from the tolerance range, the timing measurement unit 180 outputs corresponding deviation information to the servo control unit 182. The servo control unit 182 outputs a command value based on the deviation information to each servo drive circuit unit of the motor Mp (the motor Mp rotationally drives the polygon mirror PM in each scanning unit U1, U3, and U5). Each servo drive circuit section of the motor Mp consists of a feedback circuit section FBC and a servo drive circuit (amplifier) SCC. The feedback circuit section FBC receives an up/down pulse signal (hereinafter referred to as an encoder signal) from an encoder EN mounted on the rotating shaft of the motor Mp and outputs a speed signal corresponding to the rotational speed of the polygon mirror PM. The servo drive circuit (amplifier) SCC receives a command value from the servo control device 182 and the speed signal from the feedback circuit section FBC and drives the motor Mp to a rotational speed corresponding to the command value. The servo drive circuit section (feedback circuit section FBC, servo drive circuit SCC), the timing measurement section 180, and the servo control device 182 form part of the control device 18.

In this second embodiment, each polygon mirror PM in the three scanning units U1, U3, and U5 needs to maintain a certain phase difference in its rotation angle position and rotate at the same speed. In order to achieve this goal, the timing measurement unit 180 inputs the origin signals SZ1, SZ3, and SZ5, for example, and performs measurements as shown in the timing diagram of Figure 15.

Figure 15 schematically illustrates various signal waveforms generated when the three polygon mirrors PM rotate with phase differences within a predetermined tolerance range with respect to the rotation angle. Immediately after each polygon mirror PM rotates, the relative phase differences of the origin signals SZ1, SZ3, and SZ5 differ. However, the timing measurement unit 180 generates the other origin signals SZ3 and SZ5 at the same frequency (period) as origin signal SZ1, using origin signal SZ1 as a reference. The timing measurement unit 180 uses the equal time intervals Ts1, Ts2, and Ts3 between the three origin signals SZ1, SZ3, and SZ5 as a reference value and measures correction information corresponding to the error relative to this value. The timing measurement unit 180 outputs this correction information to the servo control unit 182, thereby performing servo control on the motors Mp of the scanning units U3 and U5. As shown in Figure 15, the generation timing of the three origin signals SZ1, SZ3, and SZ5 is controlled to stabilize at Ts1 = Ts2 = Ts3.

After the generation timing of origin signals SZ1, SZ3, and SZ5 stabilizes, the timing measurement unit 180 outputs drawing enable (On) signals SPP1, SPP3, and SPP5 to the selection optical elements 50, 58, and 66 shown in Figures 11 to 13 , respectively. Drawing enable (On) signals SPP1, SPP3, and SPP5 only modulate the corresponding selection optical elements 50, 58, and 66 (switching the light deflection) during their H-level periods. Since the three origin signals SZ1, SZ3, and SZ5 maintain a certain phase difference (here, one-third of the period of origin signal SZ1) after stabilization, the rising (L→H) phases of the drawing enable signals SPP1, SPP3, and SPP5 also have a certain phase difference. These drawing enable signals SPP1, SPP3, and SPP5 correspond to the drive signals (high-frequency signals) used to switch the selection optical elements 50, 58, and 66.

The timing of the falling (H to L) phase of drawing enable signals SPP1, SPP3, and SPP5 is set by measuring the clock signal CLK used to turn the spotlight on and off within each drawing line SL1, SL3, and SL5 using a counter within the timing measurement unit 180. This clock signal CLK controls the on/off timing of the drawing optical element 150 (or the drawing optical element 106 in FIG3 ) and is determined by factors such as the length of the drawing line SLn (SL1, SL3, and SL5), the size of the spotlight SP on the substrate FS, and the scanning speed Vs of the spotlight SP. For example, if the drawing line length is 30 mm, the size (diameter) of the spotlight SP is 6 μm, and the spotlight SP is turned on and off with each spotlight SP overlapping by 3 μm in the scanning direction, the drawing enable signals SPP1, SPP3, and SPP5 can be set to the falling (H to L) phase after the counter within the timing measurement unit 180 has counted the clock signal CLK to 10,000 (30 mm/3 μm) times.

Furthermore, if the polygon mirror PM has ten reflective surfaces and its rotational speed is Vp (rpm), the frequency of each origin signal SZ1, SZ3, and SZ5 is 10 Vp/60 (Hz). Therefore, when the time intervals are stabilized at Ts1 = Ts2 = Ts3, the time interval Ts1 is 60/(30 Vp) seconds. As an example, if the reference rotational speed Vp of the polygon mirror PM is 8000 rpm, the time interval Ts1 is 60/(30·8000) seconds = 250 μs.

As shown in Figure 15 , the on-time (H-level duration) Toa of the drawing enable signals SPP1, SPP3, and SPP5 is the period (projection period) during which the light beam (laser) LB from the polygon mirror PM is projected onto the substrate FS as a point light. However, it must be set shorter than the time interval Ts1. For example, if the on-time Toa is set to 200 μs, the frequency of the clock signal CLK used to count 10,000 times during this period becomes 10,000/200 = 50 MHz. Synchronously with this clock signal CLK, the drawing bit string data Sdw or serial data DLn (e.g., 10,000 bits) corresponding to the drawing line SLn generated from the pattern data ("0" or "1" on the bitmap) is output to the drawing optical element 150. In addition, as shown in Figure 3, in a structure in which a drawing optical element 106 is provided in each of the scanning units U1, U3, and U5, the drawing bit string data Sdw or serial data DL1 corresponding to the drawing line SL1 is sent to the drawing optical element 106 of the scanning unit U1, the drawing bit string data Sdw or serial data DL3 corresponding to the drawing line SL3 is sent to the drawing optical element 106 of the scanning unit U3, and the drawing bit string data Sdw or serial data DL5 corresponding to the drawing line SL5 is sent to the drawing optical element 106 of the scanning unit U5.

In this second embodiment, the drawing bit string data Sdw or serial data DLn generated from the pattern data corresponding to each of the three drawing lines SL1, SL3, and SL5 are supplied in sequence to the On/Off of the drawing optical element 150 in synchronization with the drawing enable signals SPP1, SPP3, and SPP5 (or the origin signals SZ1, SZ3, and SZ5).

FIG16 shows an example of a circuit for generating such drawing bit string data Sdw. This circuit includes generation circuits (pattern data generation circuits) 301, 303, and 305, and an OR circuit GT8. Generation circuit 301 includes a memory section BM1, a counter section CN1, and a gate section GT1. Generation circuit 303 includes a memory section BM3, a counter section CN3, and a gate section GT3. Generation circuit 305 includes a memory section BM5, a counter section CN5, and a gate section GT5. These generation circuits 301, 303, 305, and OR circuit GT8 constitute part of the control device 18.

The memory units BM1, BM3, and BM5 are memories for temporarily storing bitmap data (pattern data) corresponding to the pattern to be drawn and exposed by each scanning unit U1, U3, and U5. The counter units CN1, CN3, and CN5 are counters for outputting a bit string (e.g., 10,000 bits) corresponding to the next drawing line to be drawn, from the bitmap data (pattern data) stored in each memory unit BM1, BM3, and BM5, bit by bit as serial data DL1, DL3, and DL5 synchronized with the clock signal CLK, while the drawing enable signals SPP1, SPP3, and SPP5 are on.

The bitmap data within each memory unit BM1, BM3, and BM5 is shifted by the amount of data corresponding to one drawing line using an address counter (not shown). For example, in the case of memory unit BM1, this shifting occurs at the timing of the generation of the origin signal SZ3 by the scanning unit U3, which becomes the next active scan unit, after the output of the serial data DL1 corresponding to one drawing line is completed. Similarly, the bitmap data within memory unit BM3 is shifted at the timing of the generation of the origin signal SZ5 by the scanning unit U5, which becomes the next active scan unit, after the output of the serial data DL3 is completed. The bitmap data within memory unit BM5 is shifted at the timing of the generation of the origin signal SZ1 by the scanning unit U1, which becomes the next active scan unit, after the output of the serial data DL5 is completed.

The serial data DL1, DL3, and DL5 generated in this manner pass through gates GT1, GT3, and GT5, which are open while the drawing enable signals SPP1, SPP3, and SPP5 are on, and are applied to a three-input OR circuit GT8. OR circuit GT8 outputs the bit data sequence, synthesized repeatedly in the order of serial data DL1 → DL3 → DL5 → DL1, etc., as drawing bit string data Sdw, which is used to turn on/off the drawing optical element 150. Furthermore, as shown in FIG3 , in a configuration where each of the scanning units U1, U3, and U5 includes an drawing optical element 106, the serial data DL1 output from gate GT1 is transmitted to the drawing optical element 106 within scanning unit U1, the serial data DL3 output from gate GT3 is transmitted to the drawing optical element 106 within scanning unit U3, and the serial data DL5 output from gate GT5 is transmitted to the drawing optical element 106 within scanning unit U5.

As described above, the on/off switching of the drawing optical element 150 (or 106) needs to respond to a high-speed clock signal CLK (e.g., 50 MHz). However, the selection optical elements 50, 58, and 66 only need to be turned on and off in synchronization with the drawing enable signals SPP1, SPP3, and SPP5 (or origin signals SZ1, SZ3, and SZ5). In the previous numerical example, the response frequency can be approximately 10 kHz, since the time interval Toa (or Ts1) is 200 μs, allowing the use of inexpensive elements with high transmittance. Furthermore, if the frequency of the clock signal CLK counted by the counter in the timing measurement unit 180 or by the counter units CN1, CN3, and CN5 in FIG. 16 is denoted as Fcc, and the fundamental frequency of the pulse oscillation of the light beam LB from the light source device 14 is denoted as Fs, then n can be set to an integer greater than 1 (preferably n ≥ 2) and set so as to satisfy the relationship n·Fcc = Fs.

While the above descriptions use the operation of the light-introducing optical system 40a and the multiple scanning units U1, U3, and U5 in Figure 13, as well as the drawing timing of each scanning unit U1, U3, and U5 in Figures 14 to 16, the same applies to the light-introducing optical system 40b and the multiple scanning units U2, U4, and U6. Simply put, the scanning unit Un on which the light beam LB is incident is sequentially switched, for example, in the order of scanning unit U2, scanning unit U4, scanning unit U6, and scanning unit U2. Therefore, the control device 18 similarly switches the pattern data that determines the on/off signal sent to the drawing optical element 150 of the light-introducing optical system 40b in the order of pattern data for scanning unit U2, pattern data for scanning unit U4, pattern data for scanning unit U6, and pattern data for scanning unit U2. Furthermore, the control device 18 controls the drawing optical element 150 of the light-introducing optical system 40b based on the pattern data obtained through the sequential switching. Alternatively, a circuit configuration such as that shown in FIG16 is used to generate drawing bit string data Sdw synthesized from pattern data corresponding to three drawing lines and supply the data to the drawing optical element 150. Thus, each scanning unit U2, U4, and U6 can draw a pattern based on the pattern data on the substrate FS by irradiating the substrate FS with an intensity-modulated light beam LB along the drawing lines SL2, SL4, and SL6.

In addition to the effects of the first embodiment, the second embodiment described above also provides the following advantages. Specifically, a single image drawing optical element 150 is provided within the light introduction optical system 40a, positioned closer to the light source device 14a than the primary selection optical element 50. This single image drawing optical element 150 modulates the intensity of the light beams LB1, LB3, and LB5 emitted from the multiple scanning units U1, U3, and U5 onto the substrate FS according to a pattern. Similarly, a single image drawing optical element 150 is provided within the light introduction optical system 40b, positioned closer to the light source device 14b than the primary selection optical element 50. This single image drawing optical element 150 modulates the intensity of the light beams LB2, LB4, and LB6 emitted from the multiple scanning units U2, U4, and U6 onto the substrate FS according to a pattern. This reduces the number of acousto-optic modulation elements and reduces costs.

Furthermore, in the second embodiment described above, the drawing head 16 is described as one that divides the light beam LB into three beams. However, as described in the modified example of the first embodiment described above, a drawing head 16 that divides the light beam LB into five beams may also be used (see Figures 9 and 10). In the embodiments of Figures 9 and 10, there is only one light source device 14, and therefore, there is also only one drawing optical element 150.

[Modification of the Second Embodiment]

The second embodiment described above can be modified as follows. In the second embodiment described above, the optical element 150 for drawing is provided in the light introduction optical system 40a, 40b as a light modulator for drawing. However, in this modification, the optical element 150 for drawing is replaced with a light modulator for drawing in the light source device 14 (14a, 14b). In addition, the same figure marks are marked on the same structures as the second embodiment described above or the figures are omitted, and only the different parts are described. In addition, the light source devices provided with the light modulator for drawing in the light source devices 14a, 14b are respectively referred to as light source devices 14A and 14B. Since the light source device 14A and the light source device 14B have the same structure, only the light source device 14A will be described.

FIG17 is a diagram showing the structure of a light source device (pulse light source device, laser light source device) 14A according to this modification. Light source device 14A, which is a fiber laser device, includes a DFB semiconductor laser element 200, a DFB semiconductor laser element 202, a polarization beam splitter 204, an electro-optical element 206 serving as an image forming light modulator, a drive circuit 206a for electro-optical element 206, a polarization beam splitter 208, an absorber 210, an excitation light source 212, a combiner 214, a fiber optical amplifier 216, a wavelength conversion optical element 218, a wavelength conversion optical element 220, a plurality of lens elements GL, and a control circuit 222 including a clock generator 222a.

The DFB semiconductor laser element (first solid-state laser element, first semiconductor laser light source) 200 generates steep or sharp pulsed seed light (laser light) S1 at a predetermined frequency (oscillation frequency, fundamental frequency) Fs. The DFB semiconductor laser element (second solid-state laser element, second semiconductor laser light source) 202 generates gentle (temporally wide) pulsed seed light (laser light) S2 at the predetermined frequency Fs. The energy of a single pulse of seed light S1 generated by the DFB semiconductor laser element 200 and a single pulse of seed light S2 generated by the DFB semiconductor laser element 202 are approximately the same, but their polarization states differ, with the peak intensity of seed light S1 being higher. In this modified example, the polarization state of seed light S1 generated by the DFB semiconductor laser element 200 is S-polarization, while the polarization state of seed light S2 generated by the DFB semiconductor laser element 202 is P-polarization. The DFB semiconductor laser elements 200 and 202 are controlled to emit seed light S1 and S2 at an oscillation frequency Fs in response to a clock signal LTC (predetermined frequency Fs) generated by a clock generator 222a and electrically controlled by a control circuit 222. The control circuit 222 is controlled by the control device 18.

Furthermore, this clock signal LTC serves as the fundamental frequency of the clock signal CLK supplied to each of the counter units CN1, CN3, and CN5 shown in FIG16 . Clock signal LTC is divided by n (preferably, n is an integer greater than or equal to 2) to generate clock signal CLK. Furthermore, clock generator 222a also has the function of adjusting the fundamental frequency Fs of clock signal LTC by ±ΔF, that is, fine-tuning the time interval of the pulse oscillations of light beam LB. Thus, for example, even if the scanning speed Vs of spot light SP fluctuates slightly, the size of the pattern drawn within the drawing line range (drawing magnification) can be precisely maintained by fine-tuning fundamental frequency Fs.

Polarization beam splitter 204 transmits S-polarized light and reflects P-polarized light, guiding seed light S1 generated by DFB semiconductor laser element 200 and seed light S2 generated by DFB semiconductor laser element 202 toward electro-optical element 206. Specifically, polarization beam splitter 204 transmits S-polarized seed light S1 generated by DFB semiconductor laser element 200 to guide seed light S1 toward electro-optical element 206, and reflects P-polarized seed light S2 generated by DFB semiconductor laser element 202 to guide seed light S2 toward electro-optical element 206. DFB semiconductor laser elements 200 and 202 and polarization beam splitter 204 constitute a laser light source unit (light source unit) 205 that generates seed light S1 and S2.

The electro-optical element 206 is transmissive to the seed lights S1 and S2 and, for example, utilizes an electro-optical modulator (EOM). The EOM switches the polarization state of the seed lights S1 and S2 passing through the polarization beam splitter 204 via a driver circuit 206a in response to the on/off state (high/low) of the depicted bit string data Sdw (or serial data DLn) shown in FIG. 16 . The seed lights S1 and S2 from the DFB semiconductor laser element 200 and the DFB semiconductor laser element 202, respectively, have long wavelengths exceeding 800 nm. Therefore, an electro-optical element with a polarization state switching response of approximately GHz can be used as the electro-optical element 206.

When the pixel data of one bit of the drawing bit string data Sdw (or serial data DLn) input to the driver circuit 206a is in the off state (low, "0"), the electro-optical element 206 does not change the polarization state of the incident seed light S1 or S2 and directly guides it to the polarization beam splitter 208. On the other hand, when the drawing bit string data Sdw (or serial data DLn) input to the driver circuit 206a is in the on state (high, "1"), the electro-optical element 206 changes the polarization state of the incident seed light S1 or S2 (changes the polarization direction by 90 degrees) and then guides it to the polarization beam splitter 208. In this way, by driving the electro-optical element 206, when the pixel data of the drawing bit string data Sdw (or serial data DLn) is in the on state (high), the electro-optical element 206 converts the S-polarized seed light S1 into the P-polarized seed light S1 and converts the P-polarized seed light S2 into the S-polarized seed light S2.

The polarization beam splitter 208 transmits P-polarized light and guides it to the combiner 214 via the lens element GL, while reflecting S-polarized light and guiding it to the absorber 210. The excitation light source 212 generates excitation light, which is guided to the combiner 214 via the optical fiber 212a. The combiner 214 combines the seed light irradiated from the polarization beam splitter 208 with the excitation light and outputs it to the fiber optical amplifier (optical amplifier) 216. The fiber optical amplifier 216 is doped with a laser medium that can be excited by the excitation light. Therefore, in the fiber optical amplifier 216, through which the synthesized seed light and excitation light are transmitted, the seed light is amplified by exciting the laser medium with the excitation light. Rare earth elements such as erbium (Er), ytterbium (Yb), and thulium (Tm) are used as the laser medium doped in the fiber optical amplifier 216. The amplified seed light is emitted from the emission end 216 a of the optical fiber amplifier 216 at a predetermined divergence angle, converged or collimated by the lens element GL, and then enters the wavelength conversion optical element 218 .

The wavelength conversion optical element (first wavelength conversion optical element) 218 converts the incident seed light (wavelength λ) into a second harmonic with a wavelength of λ/2 through second harmonic generation (SHG). A PPLN (Periodically Poled LiNbO ₃ : Periodically Poled Lithium Niobate) crystal, which is a quasi-phase matching (QPM) crystal, is suitable for use as the wavelength conversion optical element 218. Alternatively, a PPLT (Periodically Poled LiTaO ₃ : Periodically Poled Lithium Tantalate) crystal can also be used.

The wavelength conversion optical element (second wavelength conversion optical element) 220 generates a third harmonic with a wavelength of λ/3 by sum frequency generation (SFG) of the second harmonic (wavelength λ/2) converted by the wavelength conversion optical element 218 and the seed light (wavelength λ) that remains without conversion by the wavelength conversion optical element 218. This third harmonic becomes ultraviolet light (light beam LB) with a peak wavelength in the wavelength range below 370 nm.

As described above, in a configuration where the drawing bit string data Sdw (or DLn) sent from the pattern data generation circuit shown in FIG16 is applied to the electro-optical element 206 in FIG17 , when the pixel data for one bit of the drawing bit string data Sdw (or DLn) is in the Off state (Low, "0"), the electro-optical element 206 does not change the polarization state of the incident seed light S1 or S2 but instead directly guides it to the polarization beam splitter 208. Therefore, the seed light that passes through the polarization beam splitter 208 becomes the seed light S2 from the DFB semiconductor laser element 202. Consequently, the light beam LB ultimately outputted from the light source device 14A has the same oscillation curve (temporal characteristics) as the seed light S2 from the DFB semiconductor laser element 202. Specifically, in this case, the light beam LB exhibits a low pulse peak intensity and a temporally broad relaxation characteristic. The fiber optical amplifier 216 has low amplification efficiency for the seed light S2 with such a low peak intensity, so the light beam LB outputted from the light source device 14A is not amplified to the energy required for exposure. Therefore, in this case, from the perspective of exposure, the result is essentially the same as when the light source device 14A does not emit the light beam LB. That is, the intensity of the point light SP irradiated onto the substrate FS becomes low. However, during the period when the pattern is not drawn along each drawing line SLn (SL1 to SL6) (non-projection period, non-exposure period), even if the light beam LB from the ultraviolet region of the seed light S2 is of a small intensity, it will continue to be radiated. Therefore, in the case where the drawing lines SLn (SL1 to SL6) are continuously in the same position on the substrate FS for a long time (for example, an emergency stop of the substrate FS due to a failure of the conveying system), it is sufficient to set a movable shutter on the emission window of the light beam LB of the light source device 14A and close the emission window.

On the other hand, when the pixel data representing one bit of the bit string data Sdw (or DLn) applied to the electro-optical element 206 of FIG. 17 is in the On state (High, "1"), the electro-optical element 206 changes the polarization state of the incident seed light S1 or S2 and guides it toward the polarization beam splitter 208. Therefore, the seed light transmitted through the polarization beam splitter 208 becomes the seed light S1 from the DFB semiconductor laser element 200. Therefore, the light beam LB output from the light source device 14A is generated by the seed light S1 from the DFB semiconductor laser element 200. Since the seed light S1 from the DFB semiconductor laser element 200 has a high peak intensity, it is efficiently amplified by the fiber optical amplifier 216, and the light beam LB output from the light source device 14A has the energy required for exposing the substrate FS. In other words, the intensity of the point light SP irradiated onto the substrate FS becomes high.

Since the electro-optical element 206 serving as a drawing light modulator is provided within the light source device 14A, the same effects as those of the second embodiment can be achieved by controlling the electro-optical element 206, as in the case of controlling the drawing optical element 150 in the second embodiment. Specifically, by switching (driving) the electro-optical element 206 on and off based on the pattern data of the scanning unit Un (or the drawing bit string data Sdw in Figures 15 and 16 ) on which the light beam LB is incident, the intensity of the light beam LB incident on the primary selection optical element 50, i.e., the intensity of the spot light SP of the light beam LB irradiated onto the substrate FS by each scanning unit Un (U1 to U6), can be modulated according to the pattern to be drawn.

Furthermore, in the configuration of FIG17 , it is also conceivable to omit the DFB semiconductor laser element 202 and polarization beam splitter 204, and instead guide only the seed light S1 pulse train (burst) from the DFB semiconductor laser element 200 to the fiber optical amplifier 216 in a wave-like manner, by switching the electro-optical element 206 based on pattern data (drawing data). However, with this configuration, the incidence periodicity of the seed light S1 into the fiber optical amplifier 216 is significantly disrupted depending on the pattern being drawn. Specifically, if the seed light S1 from the DFB semiconductor laser element 202 continues to be incident on the fiber optical amplifier 216 and then enters the fiber optical amplifier 216, the seed light S1 immediately after entry is amplified at a higher amplification factor than normal, resulting in a problem in which the fiber optical amplifier 216 generates a beam having an intensity exceeding a predetermined level. Therefore, in this modification, as a preferred method, this problem is solved by allowing seed light S2 (broad pulse light with low peak intensity) from the DFB semiconductor laser element 202 to be incident on the optical fiber amplifier 216 during the period when the seed light S1 is not incident on the optical fiber amplifier 216.

Furthermore, while switching the electro-optical element 206, the DFB semiconductor laser elements 200 and 202 can also be driven based on pattern data (depicting bit string data Sdw or serial data DLn). Specifically, the control circuit 222 controls the DFB semiconductor laser elements 200 and 202 based on the pattern data (depicting bit string data Sdw or DLn) to selectively generate seed light S1 or S2, which oscillates in a pulsed manner at a predetermined frequency Fs. In this case, the polarization beam splitters 204 and 208, the electro-optical element 206, and the absorber 210 are not required. Either of the seed light S1 or S2, which is selectively pulsed from one of the DFB semiconductor laser elements 200 and 202, is directly incident on the combiner 214. At this time, the control circuit 222 controls the driving of each DFB semiconductor laser element 200, 202 so that the seed light S1 from the DFB semiconductor laser element 200 and the seed light S2 from the DFB semiconductor laser element 202 do not simultaneously enter the optical fiber amplifier 216. Specifically, when the substrate FS is irradiated with the spot light SP of each light beam LBn, the DFB semiconductor laser element 200 is controlled so that only the seed light S1 enters the optical fiber amplifier 216. Furthermore, when the substrate FS is not irradiated with the spot light SP of the light beam LBn (the intensity of the spot light SP is extremely low), the DFB semiconductor laser element 202 is controlled so that only the seed light S2 enters the optical fiber amplifier 216. In this manner, whether or not the substrate FS is irradiated with the light beam LBn is determined based on the pixel data (high/low) of the pattern data (depicting H or L of the bit string data Sdw). In this case, the polarization states of both the seed lights S1 and S2 may be P-polarized.

As described above, also in this modification, the number of acousto-optic modulation elements can be reduced, thereby reducing the cost.

Furthermore, the light source devices 14A and 14B of this modification can also be used in place of the light source devices 14a and 14b of the first embodiment. In this case, the output timing of the seed light S1 from the DFB semiconductor laser element 200 outputted from the light source devices 14A and 14B and the switching of the drawing optical elements 106 of the scanning units U1 to U6 can be controlled based on the pattern data (drawing bit string data Sdw).

[Third embodiment]

Next, the third embodiment will be described with reference to FIG18 . This embodiment presupposes the use of the light source devices 14A (see FIG17 ) and 14B described in the modified example of the second embodiment. However, to adapt to the third embodiment, the clock generator 222a within the control circuit 222 of the light source device 14A in FIG17 has the function of locally (discretely) expanding or contracting the time interval of the clock signal LTC based on the magnification correction information CMg from the control unit (control circuit 500) for rendering control shown in FIG18 . Similarly, the clock generator 222a within the control circuit 222 of the light source device 14B also has the function of locally (discretely) expanding or contracting the time interval of the clock signal LTC based on the magnification correction information CMg. The operations of the light source device 14B, the light introduction optical system 40b, and the scanning units U2, U4, and U6 are identical to those of the light source device 14A, the light introduction optical system 40a, and the scanning units U1, U3, and U5. Therefore, descriptions of the operations of the light source device 14B, the light introduction optical system 40b, and the scanning units U2, U4, and U6 will be omitted. The same reference numerals are assigned to the same components as those in the modified example of the second embodiment, or illustration is omitted, and only the differences will be described.

In FIG18 , a light beam (laser) LB from a single light source device 14A is supplied to three scanning units U1, U3, and U5, respectively, via selection optical elements 50, 58, and 66, similar to the configurations of FIG12 and FIG13 . Selective optical elements 50, 58, and 66 selectively deflect (switches) light beam LB in response to the drawing enable (On) signals SPP1, SPP3, and SPP5 described in FIG14 and FIG15 , directing light beam LB to one of scanning units U1, U3, and U5. Furthermore, as previously described, during periods when pattern drawing along each drawing line is not being performed (non-projection periods), light beam LB from seed light S2 in the ultraviolet region continues to be emitted, even at a low intensity. To prevent situations in which each drawing line may illuminate the same position on substrate FS for an extended period, a movable shutter SST is provided on the light beam LB emission window of light source device 14A.

As shown in FIG14 , origin signals SZ1, SZ3, and SZ5 from origin sensors OP1, OP3, and OP5 of the respective scanning units U1, U3, and U5 are supplied to generation circuits (pattern data generation circuits) 301, 303, and 305, which generate pattern data for each scanning unit U1, U3, and U5. Generation circuit 301 includes gate unit GT1, memory unit BM1, and counter unit CN1, as shown in FIG16 . Counter unit CN1 is configured to count clock signal CLK1 generated using clock signal LTC output from control circuit 222 (clock generator 222a) of light source device 14A as its base frequency.

Similarly, the generating circuit 303 includes the gate part GT3, the memory part BM3, the counter part CN3, etc. in Figure 16, and the counter part CN3 is configured to count the clock signal CLK3 generated with the clock signal LTC as the base frequency. The generating circuit 305 includes the gate part GT5, the memory part BM5, the counter part CN5, etc. in Figure 16, and the counter part CN5 is configured to count the clock signal CLK5 generated with the clock signal LTC as the base frequency.

These clock signals CLK1, CLK3, and CLK5 are generated by the control circuit 500, which functions as an interface between the generation circuits 301, 303, and 305 and the light source device 14A, by dividing the clock signal LTC by 1/n (n being an integer greater than or equal to 2). The supply of these clock signals CLK1, CLK3, and CLK5 to the counter units CN1, CN3, and CN5 is limited to one of the clock signals in response to the drawing enable (On) signals SPP1, SPP3, and SPP5 (see FIG. 15 ). That is, when the drawing enable signal SPP1 is on (high), only the clock signal CLK1 obtained by dividing the clock signal LTC into 1/n is supplied to the counter section CN1. When the drawing enable signal SPP3 is on (high), only the clock signal CLK3 obtained by dividing the clock signal LTC into 1/n is supplied to the counter section CN3. When the drawing enable signal SPP5 is on (high), only the clock signal CLK5 obtained by dividing the clock signal LTC into 1/n is supplied to the counter section CN5.

Thus, the serial data DL1, DL3, and DL5 sequentially output from the respective generation circuits 301, 303, and 305 are added together via the gates GT1, GT3, and GT5, respectively, by a three-input OR circuit GT8 (see FIG. 16 ) provided within the control circuit 500, resulting in the drawing bit string data Sdw, which is then supplied to the electro-optical element 206 within the light source device 14A. The generation circuits 301, 303, and 305 and the control circuit 500 constitute a portion of the control device 18.

The above configuration is essentially the same as the method for utilizing the light source device 14A described using FIG17 . However, in this embodiment, a function is provided for individually fine-tuning the drawing magnification in the dot scanning direction (Y direction) of the pattern drawn by the drawing lines (scanning lines) SL1, SL3, and SL5 of each of the three scanning units U1, U3, and U5. To implement this function, in this embodiment, memory units BM1a, BM3a, and BM5a are provided for each scanning unit U1, U3, and U5 to temporarily store information mg1, mg3, and mg5 related to the correction amount of the drawing magnification. While these memory units BM1a, BM3a, and BM5a are illustrated as independent components in FIG18 , they may also be part of the memory units BM1, BM3, and BM5 provided in the generation circuits 301, 303, and 305, respectively. This correction amount information mg1, mg3, and mg5 also constitutes part of the drawing information.

The information mg1, mg3, and mg5 related to the correction amount corresponds to, for example, the ratio (ppm) by which the Y-direction dimension of the pattern drawn by each drawing line SL1, SL3, and SL5 is expanded or contracted. For example, if the length of the Y-direction area that can be drawn by each drawing line SL1, SL3, and SL5 is set to 30 mm, and a ±200 ppm (equivalent to ±6 μm) expansion or contraction is desired, the information mg1, mg3, and mg5 are set to a value of ±200. Alternatively, the information mg1, mg3, and mg5 can be set directly by the expansion or contraction amount (±ρμm) rather than by a ratio. Furthermore, the information mg1, mg3, and mg5 can be reset sequentially based on the pattern data (serial data DLn) for a single line along each of the drawing lines SL1, SL3, and SL5, or can be reset based on the pattern data (serial data DLn) for multiple lines. As described above, in this embodiment, while the substrate FS is being transported in the X direction (longitudinal direction) while pattern drawing is being performed along each of the drawing lines SL1, SL3, and SL5, the drawing magnification in the Y direction can be dynamically changed. This allows for the suppression of degradation in drawing position accuracy caused by the deformation and in-plane skew of the substrate FS. Furthermore, during overlay exposure, deformation of the already formed underlying pattern can be accounted for, significantly improving overlay accuracy.

FIG19 is a timing diagram showing the signal states of each part and the oscillation state of the light beam LB in the drawing device shown in FIG18 , representatively when a standard pattern drawing is performed based on the scanning unit U1. In FIG19 , the two-dimensional matrix Gm shows the bit pattern PP of the pattern data to be drawn, and one grid (1 pixel unit) on the substrate FS is set to, for example, a size Py of 3 μm in the Y direction and a size Px of 3 μm in the X direction. In addition, in FIG19 , SL1-1, SL1-2, SL1-3, ... SL1-6 indicated by arrows show drawing lines drawn in sequence by the drawing line SL1 as the substrate FS moves in the X direction (sub-scanning in the long side direction), and the conveying speed of the substrate FS is set so that the interval in the X direction of each drawing line SL1-1, SL1-2, SL1-3, ... SL1-6 becomes, for example, 1/2 of the size Px (3 μm) of 1 pixel unit.

Moreover, the size (spot size) of the point light SP projected onto the substrate FS in the XY direction is set to the same degree as or slightly larger than a pixel unit. Thus, the size of the point light SP is set to about 3 to 4 μm as an effective diameter (the width of 1/e ² of the Gaussian distribution, or the half-value full width of the peak intensity). When the point light SP is continuously projected along the drawing line SL1, the oscillation frequency Fs (pulse time interval) of the light beam LB and the scanning speed Vs of the point light SP achieved by the polygonal mirror PM are set in a manner such that they overlap by, for example, 1/2 of the effective diameter of the point light SP. That is, if the seed light emitted from the polarization beam splitter 208 in the light source device 14A shown in FIG17 is set to a light beam Lse ( FIG18 ), then the seed light beam Lse is emitted as shown in FIG19 in response to each clock pulse of the clock signal LTC output from the control circuit 222 (clock generator 222a).

The clock signal LTC and the clock signal CLK1 supplied to the counter unit CN1 in the generator circuit 301 in FIG18 are set to a frequency ratio of 1:2. When the clock signal LTC is 100 MHz, the clock signal CLK1 is set to 50 MHz by the 1/2 frequency divider of the control circuit 500 in FIG18. The frequency ratio of the clock signal LTC to the clock signal CLK1 can be any integer multiple. For example, the set frequency of the clock signal CLK1 can be reduced to 1/4, that is, to 25 MHz, and the scanning speed Vs of the spot light SP can also be reduced to half.

The drawing bit string data Sdw shown in FIG19 corresponds to the serial data DL1 output from the generating circuit 301. Here, for example, it corresponds to the pattern on the drawing line SL1-2 of the pattern PP. The electro-optical element 206 within the light source device 14A switches its polarization state in response to the drawing bit string data Sdw. Therefore, the seed light beam Lse is generated by the seed light S1 from the DFB semiconductor laser element 200 in FIG17 during the period when the drawing bit string data Sdw is in the on state (high, "1"). It is generated by the seed light S2 from the DFB semiconductor laser element 202 in FIG17 during the period when the drawing bit string data Sdw is in the off state (low, "0"). The drawing exposure operation of the scanning unit U1 shown in FIG19 is similar to that of the other scanning units U2 to U6.

In addition, when a driving circuit is provided in the control circuit 222 of the light source device 14A (the driving circuit generates seed light S1 (rapidly rising and falling pulse light) from the DFB semiconductor laser element 200 in response to the clock signal LTC during the period when the bit string data Sdw is depicted in the On state (high, "1"), and generates seed light S2 (broad pulse light) from the DFB semiconductor laser element 202 in response to the clock signal LTC during the period when the bit string data Sdw is depicted in the Off state (low, "0")), the electro-optical element 206 shown in Figures 17 and 18, the polarization beam splitter 208 shown in Figure 17, and the absorber 210 can be omitted.

As described above, each pulse of the seed light beam Lse is output in response to each clock pulse of the clock signal LTC generated by the clock generator 222a shown in FIG17 . Therefore, in this embodiment, a circuit structure for locally increasing or decreasing the time (period) between pulses of the clock signal LTC is provided within the clock generator 222a. This circuit structure includes a reference (standard) clock generator, which serves as the source of the clock signal LTC, a frequency division counter circuit, and a variable delay circuit.

FIG20 is a timing chart showing the relationship between the reference clock signal TC0 and the clock signal LTC from the reference clock generator within clock generator 222a. This chart illustrates a state where no correction based on the magnification correction information CMg shown in FIG17 and FIG18 is being performed. The variable delay circuit within clock generator 222a constantly delays the reference clock signal TC0, generated at a fixed frequency Fs (fixed time Td0), by a delay time DT0 corresponding to a preset value and outputs it as the clock signal LTC. Therefore, for example, if the reference clock signal TC0 is 100 MHz (Td0 = 10 nS), the clock signal LTC continues to be generated at 100 MHz (Td0 = 10 nS) as long as the preset value (delay time DT0) remains unchanged.

Therefore, the following configuration is employed: the frequency division counter circuit within clock generator 222a counts the reference clock signal TC0. When the count reaches a predetermined value Nv, the preset value set for the variable delay circuit is changed by a fixed amount. This is illustrated by the time chart of Figure 21. In Figure 21, before the reference clock signal TC0 reaches Nv by the frequency division counter circuit, the preset value set for the variable delay circuit is the delay time DT0. Thereafter, when the frequency division counter circuit reaches Nv due to a single clock pulse Kn of the reference clock signal TC0, the preset value set for the variable delay circuit is immediately changed to the delay time DT1. Therefore, each clock pulse (K'n+1 and later) of the clock signal LTC generated based on the clock pulse Kn+1 following the clock pulse Kn of the reference clock signal TC0 is uniformly generated with a delay time DT1.

Thus, only when the preset value set for the variable delay circuit is changed by a certain amount—that is, only when the time interval Td1 is changed between clock pulses K'n and K'n+1 of the clock signal LTC—does the time interval between the clock pulses of the clock signal LTC thereafter become Td0. In Figure 21 , the delay time DT1 is increased compared to the delay time DT0, and the time between the two clock pulses of the clock signal LTC is increased compared to Td0, but it can also be reduced. Furthermore, after the frequency division counter circuit counts the reference clock signal TC0 to Nv, it returns to zero and begins counting again to Nv.

If the initial value of the preset value set for the variable delay circuit is set to delay time DT0, the amount of delay time change is set to ±ΔDh, the number of times the frequency division counter circuit is reset to zero is set to Nz, and the delay time that is sequentially set to the preset value for the variable delay circuit each time the frequency division counter circuit counts to Nv (each time it is reset to zero) is set to DTm, then the delay time DTm is set to the relationship DTm = DT0 + Nz (±ΔDh). Therefore, as shown in Figure 21, the delay time DT1 set during the period when the number of resets to zero is 1 (m=1) is DTm = DT1 = DT0 ±ΔDh. The delay time DT2 set after the next reset to zero (Nz=2, m=2) is DTm = DT2 = DT0 + 2 (±ΔDh). Therefore, the amount of delay time change ±ΔDh corresponds to the difference in time Td1 between clock pulses K'n and K'n+1 of the clock signal LTC relative to the reference time Td0.

As described above, the operation of varying the time interval between specific two clock pulses of the clock signal LTC is discretely performed at multiple locations along the entire length of a plotted line (SL1-SL6) based on the predetermined value Nv set for the frequency division counter circuit. Figure 22 illustrates this. Figure 22 shows multiple locations along the entire length of plotted line SL1 where the count value of the frequency division counter circuit returns to zero each time it reaches the predetermined value Nv, as correction points CPP. At each of these correction points CPP, only the time interval between specific two clock pulses of the clock signal LTC is stretched or shortened by ±ΔDh relative to time Td0.

Therefore, if the reference clock signal TC0 is set to 100 MHz (Td0 = 10 nS), the effective size of the spot light SP in the main scanning direction is set to 3 μm, and the length of the drawing line SL1 (the same applies to SL2 to SL6) is set to 30 mm, and two consecutive pulses of light from the light beam LB projected onto the substrate FS overlap the spot light SP by approximately half (1.5 μm) in the main scanning direction, the number of reference clock signal TC0 clocks generated within the length of the drawing line SL1 is 20,000. Furthermore, the change in delay time ΔDh relative to the reference time interval Td0 is very small, for example, set to approximately 2%. Under these conditions, if the pattern drawn along the drawing line SL1 is expanded or contracted by 150 ppm in the main scanning direction (Y direction), 150 ppm of the 30 mm length of the drawing line SL1 is equivalent to 4.5 μm. Information regarding the ratio of the drawing magnification (150 ppm) or the effective size (4.5 μm) is stored as information mg1 in the memory unit BM1a in FIG18.

Therefore, the number of correction points CPP (Figure 22) at which the time ΔDh is expanded or contracted relative to the time Td0 in the 20,000 clock pulse train of the clock signal LTC becomes 4.5 μm/(1.5 μm × 2%) = 150. The maximum specified value Nv set for the frequency division counter circuit shown in Figure 22 is approximately 133, which is 20,000/150.

Furthermore, when the delay time variation ΔDh is set to 5%, the number of correction points CPP becomes 4.5 μm/(1.5 μm × 5%) = 60, and the maximum predetermined value Nv set for the frequency division counter circuit is approximately 333, which is 20,000/60. Since the delay time variation ΔDh is small, less than 10%, even if a pattern to be drawn exists at the correction point CPP, the size of the pattern is larger than the size of the spot light SP. Therefore, a drawing error caused by a slight misalignment of the spot light SP in the main scanning direction at the correction point CPP can be ignored.

The above-mentioned change in delay time ΔDh, the number of correction points CPP, the setting of the specified value Nv for the frequency division counter circuit, etc. are calculated in the control circuit 222 shown in Figure 17 based on the multiplication correction information CMg (ppm) output from the control circuit 500 of Figure 18, and the frequency division counter circuit, variable delay circuit, etc. in the clock generator 222a are set.

According to the above embodiment, the light beam LB from the light source device 14A can be supplied sequentially to, for example, the three scanning units U1, U3, and U5 in a time-division manner, allowing each scanning unit U1, U3, and U5 to perform drawing operations along the drawing lines SL1, SL3, and SL5 in sequence. Therefore, as shown in FIG18 , information mg1, mg3, and mg5 related to the correction amount of the drawing magnification can be set for each scanning unit U1, U3, and U5. Thus, even if the Y-direction expansion and contraction of the substrate FS differ, or the expansion and contraction rates of the several regions divided along the Y-direction differ, the optimal correction amount of the drawing magnification can be set for each scanning unit Un in a manner corresponding to the respective expansion and contraction rates, thereby achieving the advantage of being able to cope with nonlinear deformation of the substrate FS.

As described above, in a light source device 14A connected to a device for scanning a point light SP converged on an irradiated object (substrate FS) to draw a pattern and emitting a light beam (laser) LB that will become the point light SP, as shown in Figures 17 and 18, there are provided: a first semiconductor laser light source (200), which responds to a clock pulse (clock signal LTC) of a specified period (Td0) to generate a first pulse light (seed light S1) with a short luminous time relative to the specified period and a high peak intensity that rises and falls sharply; a second semiconductor laser light source (202), which responds to the clock pulse to generate a second pulse light (seed light S2) with a shorter luminous time than the specified period and a longer luminous time than the first pulse light (seed light S1) and a low peak intensity that is broad; and an optical fiber amplifier (216) for supplying the first pulse light (seed light S1) or the second pulse light (seed light The invention relates to a method for producing a first pulsed light (seed light S1) incident on an optical fiber amplifier (216) based on information of a pattern to be produced (drawing bit string data Sdw), and a switching device for switching the method so that a first pulsed light (seed light S1) is incident on an optical fiber amplifier when a point light SP is projected onto an irradiated object, and a second pulsed light (seed light S2) is incident on an optical fiber amplifier (216) when a point light SP is not projected onto the irradiated object. The switching device is composed of an electro-optical element (206) for selecting one of the first pulsed light (seed light S1) and the second pulsed light (seed light S2) based on information of a pattern to be produced, or a circuit for controlling the driving of a first semiconductor laser light source (200) and a second semiconductor laser light source (202) based on information of a pattern to be produced so that one of the first pulsed light (seed light S1) and the second pulsed light (seed light S2) is produced.

This third embodiment can also be applied to the aforementioned first embodiment or its variations, and the aforementioned second embodiment. Specifically, the function described in the third embodiment of the clock generator 222a within the control circuit 222 of the light source device 14A, which locally (discretely) expands or contracts the time interval of the clock signal LTC based on the magnification correction information CMg from the control unit (control circuit 500) for rendering control shown in FIG18 , can be applied to the light source device 14 of the aforementioned first embodiment or its variations, and the light source device 14 of the aforementioned second embodiment. In this case, the light source device 14 need not include the DFB semiconductor laser element 202, the polarization beam splitter 204, the electro-optical element 206, the polarization beam splitter 208, and the absorber 210. In other words, the light source device 14 may amplify the pulsed seed light S1 emitted by the DFB semiconductor laser element 200 using the fiber optical amplifier 216 and emit it as the light beam LB. In this case, since the light source device 14 does not include the electro-optical element 206 , the serial data DL1 , DL3 , and DL5 generated by the generation circuits 301 , 303 , and 305 are sent to the image drawing optical element 106 or the image drawing optical element 150 of the scanning unit Un.

[Fourth embodiment]

FIG23 is a diagram schematically illustrating the configuration of a device manufacturing system 10 according to a fourth embodiment, including an exposure apparatus EX for performing an exposure process on a substrate (irradiated object) FS. Unless otherwise specified, components identical to those of the first to third embodiments (including modifications) described above are denoted by the same reference numerals or omitted from illustration, and only the differences are described.

In this fourth embodiment, similar to the first to third embodiments (including modified examples), the exposure apparatus EX, serving as a beam scanning device, is a direct-drawing exposure apparatus that does not use a mask, i.e., a so-called raster scanning exposure apparatus. The exposure apparatus EX replaces the drawing head 16 described in the first to third embodiments (including modified examples) with a beam switching unit 20 and an exposure head 22. Furthermore, the exposure apparatus EX includes multiple alignment microscopes AMm (AM1 to AM4). Although not specifically described in the first to third embodiments (including modified examples), the exposure apparatus EX of the first to third embodiments also includes multiple alignment microscopes AMm (AM1 to AM4). Furthermore, the exposure apparatus EX of the fourth embodiment also includes a substrate transport mechanism 12, a light source device 14', and a control device 18. Furthermore, the light source device 14' of the fourth embodiment is based on the same structure as the light source device 14 (light source devices 14A and 14B) described in the modified example of the second embodiment (see FIG. 17 ). The light beam LB emitted from the light source device 14' passes through the light beam switching component 20 and enters the exposure head 22.

The light beam switching unit 20 switches the optical path of the light beam LB from the light source device 14' so that the light beam LB is incident on one of the plurality of scanning units Un (U1 to U6) constituting the exposure head 22, which performs one-dimensional scanning of the spot light SP. The light beam switching unit 20 will be described in detail later.

The exposure head 22 includes a plurality of scanning units Un (U1 to U6) into which the light beams LB are incident. The exposure head 22 draws a pattern on a portion of the substrate FS supported by the circumferential surface of the rotating drum DR through the plurality of scanning units Un (U1 to U6). The exposure head 22 is a so-called multi-beam type exposure head in which a plurality of scanning units Un (U1 to U6) of the same structure are arranged. As shown in FIG23 , the odd-numbered scanning units U1, U3, and U5 are arranged on the upstream side (-X direction side) of the conveying direction of the substrate FS relative to the center plane Poc and are arranged along the Y direction. The even-numbered scanning units U2, U4, and U6 are arranged on the downstream side (+X direction side) of the conveying direction of the substrate FS relative to the center plane Poc and are arranged along the Y direction. The odd-numbered scanning units U1, U3, and U5 are symmetrically arranged with respect to the even-numbered scanning units U2, U4, and U6 with respect to the center plane Poc. That is, in the fourth embodiment, the arrangement of the odd-numbered scanning units U1 , U3 , and U5 and the even-numbered scanning units U2 , U4 , and U6 is opposite to that described in the first to third embodiments (including modifications).

The scanning unit Un projects the light beam LB from the light source device 14' in a manner that converges into a point light SP on the irradiated surface of the substrate FS, and at the same time uses a rotating polygonal mirror PM (refer to Figure 28) to cause the point light SP to be scanned in one dimension along a prescribed straight line drawing line (scanning line) SLn on the irradiated surface of the substrate FS.

A plurality of scanning units Un (U1 to U6) are arranged in a prescribed configuration relationship. In the present fourth embodiment, the plurality of scanning units Un (U1 to U6) are arranged so that the drawing lines SLn (SL1 to SL6) of the plurality of scanning units Un (U1 to U6) are joined together in the Y direction (the width direction of the substrate FS, the main scanning direction) without being separated from each other as shown in FIG. 24 and FIG. 25 . In addition, as described in the first to third embodiments (modifications), there is a case where the light beams LB incident on each scanning unit Un (U1 to U6) are respectively represented as LB1 to LB6. The light beam LB incident on the scanning unit Un is a linearly polarized light beam (P polarization or S polarization) polarized in a prescribed direction, and in the present fourth embodiment, it is a P polarized light beam. In addition, there is also a case where the light beams LB1 to LB6 incident on each of the six scanning units U1 to U6 are represented as light beams LBn.

As shown in FIG25 , the scanning area is shared by each scanning unit Un (U1 to U6) in such a manner that all of the scanning units in the plurality of scanning units Un (U1 to U6) cover the entire range in the width direction of the exposure area W. Thus, each scanning unit Un (U1 to U6) can draw a pattern in each of the plurality of areas divided along the width direction of the substrate FS. For example, if the scanning length in the Y direction (the length of the drawing line SLn) achieved by one scanning unit Un is set to approximately 30 to 60 mm, then by arranging three scanning units, namely odd-numbered scanning units U1, U3, and U5, and three scanning units, namely even-numbered scanning units U2, U4, and U6, a total of six scanning units Un, along the Y direction, the width in the Y direction that can be drawn is expanded to approximately 180 to 360 mm. In principle, the lengths (scanning lengths, scanning widths in the main scanning direction) of the drawing lines SL1 to SL6 are the same.

As described above, each actual drawing line SLn (SL1-SL6) is set slightly shorter than the maximum length that the spot light SP can actually scan on the illuminated surface. This setting allows for fine adjustment of the position of the drawing line SLn (for example, a scanning length of 30 mm) in the main scanning direction and the drawing magnification within the maximum scanning length of the spot light SP (for example, 31 mm). The maximum scanning length of the spot light SP is primarily determined by the aperture of the fθ lens FT (see Figure 28), located after the polygon mirror (rotating polygon mirror) PM within the scanning unit Un.

Multiple drawing lines SLn (SL1-SL6) are arranged in two rows in the circumferential direction of the rotating drum DR, sandwiched between the center plane Poc. The odd-numbered drawing lines SL1, SL3, and SL5 are located on the irradiated surface of the substrate FS upstream (-X direction side) in the conveyance direction of the substrate FS relative to the center plane Poc. The even-numbered drawing lines SL2, SL4, and SL6 are located on the irradiated surface of the substrate FS downstream (+X direction side) in the conveyance direction of the substrate FS relative to the center plane Poc. The drawing lines SL1-SL6 are approximately parallel to the width direction of the substrate FS, that is, the center axis AXo of the rotating drum DR.

Drawing lines SL1, SL3, and SL5 are arranged on a straight line at predetermined intervals along the width direction (scanning direction) of the substrate FS. Drawing lines SL2, SL4, and SL6 are similarly arranged on a straight line at predetermined intervals along the width direction (scanning direction) of the substrate FS. The scanning direction of the spot light SP of the light beam LBn scanning along each of the odd-numbered drawing lines SL1, SL3, and SL5 is one-dimensional, namely, the -Y direction. The scanning direction of the spot light SP of the light beam LBn scanning along each of the even-numbered drawing lines SL2, SL4, and SL6 is one-dimensional, namely, the +Y direction.

In the fourth embodiment, a plurality of scanning units Un (U1 to U6) repeatedly scan the point light SP of the light beam LBn in a predetermined order (prescribed order). For example, when the order of the scanning units Un that scan the point light SP is U1 → U2 → U3 → U4 → U5 → U6, first, the scanning unit U1 scans the point light SP once. Then, when the scanning unit U1 finishes scanning the point light SP, the scanning unit U2 scans the point light SP once, and when the scanning finishes, the scanning unit U3 scans the point light SP once. In this way, the plurality of scanning units Un (U1 to U6) each scans the point light SP once in the prescribed order. Moreover, when the scanning unit U6 finishes scanning the point light SP, the scanning unit U1 returns to scanning the point light SP. In this way, the plurality of scanning units Un (U1 to U6) repeatedly scans the point light SP in the prescribed order.

Each scanning unit Un (U1 to U6) irradiates each light beam LBn toward the substrate FS in such a manner that each light beam LBn travels toward the center axis AXo of the rotating cylinder DR, at least in the XZ plane. Thus, the optical path (the center axis of the light beam) of the light beam LBn traveling from each scanning unit Un (U1 to U6) toward the substrate FS is coaxial (parallel) with the normal to the irradiated surface of the substrate FS in the XZ plane. In addition, each scanning unit Un (U1 to U6) irradiates the light beam LBn toward the drawing line SLn (SL1 to SL6) toward the substrate FS in such a manner that the light beam LBn is perpendicular to the irradiated surface of the substrate FS within a plane parallel to the YZ plane. That is, with respect to the main scanning direction of the point light SP on the irradiated surface, the light beam LBn (LB1 to LB6) projected onto the substrate FS is scanned in a telecentric state. Here, a line (or also called an optical axis) passing through each midpoint of the drawing line SLn (SL1 to SL6) defined by each scanning unit Un (U1 to U6) and perpendicular to the irradiated surface of the substrate FS is called the irradiation center axis Len (Le1 to Le6) (refer to Figure 24).

Each irradiation center axis Len (Le1-Le6) is a line connecting the drawing lines SL1-SL6 and the center axis AXo in the XZ plane. The irradiation center axes Le1, Le3, and Le5 of the odd-numbered scanning units U1, U3, and U5 are aligned in the same direction in the XZ plane, while the irradiation center axes Le2, Le4, and Le6 of the even-numbered scanning units U2, U4, and U6 are aligned in the same direction in the XZ plane. Furthermore, the irradiation center axes Le1, Le3, and Le5 and the irradiation center axes Le2, Le4, and Le6 are set at angles of ±θ relative to the center plane Poc in the XZ plane (see Figure 23).

The alignment microscope AMm (AM1 to AM4) shown in FIG23 is used to detect the alignment marks MKm (MK1 to MK4) formed on the substrate FS, as shown in FIG25 , and is provided with a plurality of them along the Y direction (four in the present fourth embodiment). The alignment marks MKm (MK1 to MK4) are reference marks for aligning (aligning) the prescribed pattern to be drawn in the exposure area W on the irradiated surface of the substrate FS relative to the substrate FS. The alignment microscope AMm (AM1 to AM4) detects the alignment marks MKm (MK1 to MK4) on the substrate FS supported by the circumferential surface of the rotating drum DR. The alignment microscope AMm (AM1 to AM4) is provided on the upstream side (-X direction side) of the conveying direction of the substrate FS compared to the irradiated area (the area surrounded by the drawing lines SL1 to SL6) formed on the substrate FS based on the point light SP of the light beam LBn (LB1 to LB6) from the exposure head 22.

The alignment microscope AMm (AM1 to AM4) includes: a light source for projecting illumination light for alignment onto the substrate FS, an observation optical system (including an objective lens) for obtaining a magnified image of a local area (observation area) including the alignment marks MKm (MK1 to MK4) on the surface of the substrate FS, and an imaging element such as a CCD or CMOS for capturing the magnified image with a high-speed shutter while the substrate FS moves along the conveying direction. The imaging signal (image data) ig (ig1 to ig4) captured by the alignment microscope AMm (AM1 to AM4) is sent to the control device 18. The control device 18 detects the position of the alignment marks MKm (MK1 to MK4) based on image analysis of the imaging signal ig (ig1 to ig4) and information on the rotational position of the rotating drum DR at the moment of capturing the image (based on the measured values obtained by the encoders EN1a and EN1b of the reading scale section SD shown in FIG24), thereby detecting the position of the substrate FS with high precision. In addition, the alignment illumination light is light in a wavelength band to which the photosensitive functional layer on the substrate FS has almost no sensitivity, for example, light with a wavelength of about 500 to 800 nm.

Alignment marks MK1 to MK4 are arranged around each exposure area W. A plurality of alignment marks MK1 and MK4 are formed at a certain interval DI on both sides of the width direction of the substrate FS of the exposure area W along the long side direction of the substrate FS. Alignment mark MK1 is formed on the -Y direction side of the width direction of the substrate FS, and alignment mark MK4 is formed on the +Y direction side of the width direction of the substrate FS. Such alignment marks MK1 and MK4 are configured so as to be located at the same position in the long side direction (X direction) of the substrate FS when the substrate FS is not deformed by large tension or heat treatment. Moreover, alignment marks MK2 and MK3 are formed between alignment mark MK1 and alignment mark MK4, and in the blank portions on the +X direction side and -X direction side of the exposure area W along the width direction (short side direction) of the substrate FS. Alignment marks MK2 and MK3 are formed between exposure area W and exposure area W. Alignment mark MK2 is formed on the -Y direction side of the width direction of the substrate FS, and alignment mark MK3 is formed on the +Y direction side of the substrate FS.

Moreover, the interval in the Y direction between the alignment mark MK1 arranged at the side end of the -Y direction of the substrate FS and the alignment mark MK2 of the blank part, the interval in the Y direction between the alignment mark MK2 of the blank part and the alignment mark MK3, and the interval in the Y direction between the alignment mark MK4 arranged at the side end of the +Y direction of the substrate FS and the alignment mark MK3 of the blank part are all set to the same distance. These alignment marks MKm (MK1 to MK4) can be formed together when the pattern layer of the first layer is formed. For example, when the pattern of the first layer is exposed, the pattern for the alignment mark can also be exposed around the exposure area W where the pattern is to be exposed. In addition, the alignment mark MKm can also be formed in the exposure area W. For example, it can be formed in the exposure area W along the contour of the exposure area W. In addition, when the alignment mark MKm is formed in the exposure area W, the pattern portion at a specific position or a portion of a specific shape in the pattern of the electronic device formed in the exposure area W can also be used as the alignment mark MKm.

The alignment microscope AM1 is configured to photograph the alignment mark MK1 present in the observation area (detection area) Vw1 of the objective lens. Similarly, the alignment microscopes AM2 to AM4 are configured to photograph the alignment marks MK2 to MK4 present in the observation area Vw2 to Vw4 of the objective lens. Therefore, a plurality of alignment microscopes AM1 to AM4 are arranged in the order of the alignment microscopes AM1 to AM4 from the -Y direction side of the substrate FS, corresponding to the positions of the plurality of alignment marks MK1 to MK4. The alignment microscopes AMm (AM1 to AM4) are arranged so that the distance between the exposure position (drawing lines SL1 to SL6) and the observation area Vw (Vw1 to Vw4) of the alignment microscope AMm in the X direction is shorter than the length of the exposure area W in the X direction. The number of alignment microscopes AMm arranged in the Y direction can be changed according to the number of alignment marks MKm formed in the width direction of the substrate FS. In addition, the size of the observation areas Vw1 to Vw4 on the irradiated surface of the substrate FS is set according to the size of the alignment marks MK1 to MK4 and/or the alignment accuracy (position measurement accuracy), and is about 100 to 500 μm square. In addition, although not specifically described in the first to third embodiments (including modified examples), a plurality of alignment marks MKm are also formed on the substrate FS used in the first to third embodiments.

As shown in Figure 24, at both ends of the rotating cylinder DR, there are provided scale portions SD (SDa, SDb) having scales formed in an annular shape over the entire circumferential range of the outer peripheral surface of the rotating cylinder DR. The scale portions SD (SDa, SDb) are diffraction gratings with concave or convex grid lines engraved at a certain interval (for example, 20 μm) in the circumferential direction of the outer peripheral surface of the rotating cylinder DR, constituting an incremental scale. The scale portions SD (SDa, SDb) rotate integrally with the rotating cylinder DR around the central axis AXo. In addition, a plurality of encoders (scale reading heads) ENn are provided in a manner opposite to the scale portions SD (SDa, SDb). The encoder ENn optically detects the rotational position of the rotating cylinder DR. Opposite to the scale portion SDa provided at the end portion on the -Y direction side of the rotating cylinder DR, three encoders ENn (EN1a, EN2a, EN3a) are provided. Similarly, three encoders ENn (EN1b, EN2b, EN3b) are provided opposite the scale portion SDb provided at the end portion of the rotating drum DR on the +Y direction side. Furthermore, although not specifically described in the first to third embodiments (including variations), scale portions SD (SDa, SDb) are provided at both ends of the rotating drum DR in the first to third embodiments, and a plurality of encoders En (EN1a to EN3a, EN1b to EN3b) are provided opposite thereto.

The encoders ENn (EN1a-EN3a, EN1b-EN3b) project a measuring beam toward the scale unit SD (SDa, SDb) and photoelectrically detect the reflected beam (diffracted light), outputting a detection signal as a pulse signal to the control device 18. The control device 18 counts these detection signals (pulse signals) using a counter circuit 356a (see FIG. 33 ), thereby measuring the rotational angular position and angular change of the rotating drum DR with submicron resolution. The counter circuit 356a counts the detection signals from each encoder ENn (EN1a-EN3a, EN1b-EN3b) individually. The control device 18 can also measure the transport speed of the substrate FS based on the angular change of the rotating drum DR. The counter circuit 356a, which counts the detection signals of each encoder ENn (EN1a~EN3a, EN1b~EN3b) individually, resets the count value corresponding to the encoder ENn to 0 when each encoder ENn (EN1a~EN3a, EN1b~EN3b) detects the origin mark (origin pattern) ZZ formed on a circumferential part of the scale part SDa, SDb.

Encoders EN1a and EN1b are arranged on an installation azimuth line Lx1. Installation azimuth line Lx1 is a line connecting the projection position (reading position) of the measurement beams of encoders EN1a and EN1b onto scale unit SD (SDa, SDb) with the central axis AXo in the XZ plane. Furthermore, installation azimuth line Lx1 is a line connecting the observation area Vw (Vw1 to Vw4) of each alignment microscope AMm (AM1 to AM4) with the central axis AXo in the XZ plane.

Encoders EN2a and EN2b are located upstream (-X direction) of the substrate FS conveyance direction relative to the center plane Poc, and downstream (+X direction) of the substrate FS conveyance direction relative to encoders EN1a and EN1b. Encoders EN2a and EN2b are arranged on a setting azimuth line Lx2. Setting azimuth line Lx2 is a line connecting the projection position of the measurement light beams of encoders EN2a and EN2b onto the scale unit SD (SDa, SDb) with the center axis AXo in the XZ plane. Setting azimuth line Lx2 overlaps with the irradiation center axes Le1, Le3, and Le5 at the same angular position in the XZ plane.

Encoders EN3a and EN3b are located downstream (on the +X direction) of the substrate FS transport direction relative to the center plane Poc. Encoders EN3a and EN3b are arranged on a setting azimuth line Lx3. Setting azimuth line Lx3 is a line connecting the projection positions of the measurement beams of encoders EN3a and EN3b onto the scale unit SD (SDa, SDb) with the center axis AXo in the XZ plane. Setting azimuth line Lx3 overlaps with the irradiation center axes Le2, Le4, and Le6 at the same angular position in the XZ plane.

The count values (rotational angle positions) of the detection signals from encoders EN1a and EN1b, the count values (rotational angle positions) of the detection signals from encoders EN2a and EN2b, and the count values (rotational angle positions) of the detection signals from encoders EN3a and EN3b are reset to zero at the moment each encoder ENn detects an origin mark ZZ located at a location in the rotation direction of the rotating drum DR. Therefore, when the count values based on encoders EN1a and EN1b are set to a first value (e.g., 100), and the position of the substrate FS wound around the rotating drum DR on the setting orientation line Lx1 (the positions of the observation areas Vw1 to Vw4 of the alignment microscopes AM1 to AM4) is set to the first position, when the first position on the substrate FS is conveyed to a position on the setting orientation line Lx2 (the positions of the drawing lines SL1, SL3, and SL5), the count values based on encoders EN2a and EN2b become the first value (e.g., 100). Similarly, when the first position on the substrate FS is conveyed to the position on the setting orientation line Lx3 (the position of the drawing lines SL2, SL4, and SL6), the count value based on the detection signal of the encoders EN3a and EN3b becomes the first value (for example, 100).

However, the substrate FS is wound on the inner side compared to the scale parts SDa and SDb at both ends of the rotating drum DR. In Figure 23, the radius of the outer peripheral surface of the scale part SD (SDa, SDb) from the center axis AXo is set to be smaller than the radius of the outer peripheral surface of the rotating drum DR from the center axis AXo. However, as shown in Figure 24, it can also be set so that the outer peripheral surface of the scale part SD (SDa, SDb) and the outer peripheral surface of the substrate FS wound on the rotating drum DR become coplanar. That is, it can also be set so that the radius (distance) of the outer peripheral surface of the scale part SD (SDa, SDb) from the center axis AXo is the same as the radius (distance) of the outer peripheral surface (irradiated surface) of the substrate FS wound on the rotating drum DR from the center axis AXo. Therefore, the encoder ENn (EN1a, EN1b, EN2a, EN2b, EN3a, EN3b) can detect the scale part SD (SDa, SDb) at the same radial position as the irradiated surface of the substrate FS wound on the rotating drum DR, and can reduce the Abbe error caused by the difference between the measurement position of the encoder ENn and the processing position (drawing lines SL1~SL6) in the radial direction of the rotating drum DR.

Based on the above, the control device 18 determines the start position of the drawing exposure of the exposure area W in the longitudinal direction (X direction) of the substrate FS based on the positions of the alignment marks MKm (MK1-MK4) detected by the alignment microscopes AMm (AM1-AM4) (based on the count values obtained by encoders EN1a and EN1b). At this time, the count values obtained by encoders EN1a and EN1b are set to a first value (e.g., 100). In this case, when the count values obtained by encoders EN2a and EN2b reach a first value (e.g., 100), the start position of the drawing exposure of the exposure area W in the longitudinal direction of the substrate FS is located on the drawing lines SL1, SL3, and SL5. Therefore, the scanning units U1, U3, and U5 can start scanning with the spot light SP based on the count values obtained by encoders EN2a and EN2b. Furthermore, when the count values of encoders EN3a and EN3b reach a first value (e.g., 100), the start position of the drawing exposure of the exposure area W in the longitudinal direction of the substrate FS is located on the drawing lines SL2, SL4, and SL6. Therefore, the scanning units U2, U4, and U6 can start scanning the spot light SP based on the count values of encoders EN3a and EN3b. Furthermore, although not specifically described in the first to third embodiments (including modifications), the exposure apparatus EX of the first to third embodiments also includes encoders ENn (EN1a to EN3a, EN1b to EN3b) and a scale unit SD (SDa and SDb).

Figure 26 is a structural diagram of the beam switching unit 20. The beam switching unit 20 includes multiple selection optical elements AOMn (AOM1-AOM6), multiple focusing lenses CD1-CD6, multiple reflectors M1-M12, multiple unit-side incident mirrors IM1-IM6, multiple collimating lenses CL1-CL6, and an absorber TR. The selection optical elements AOMn (AOM1-AOM6) are transmissive to the light beam LB and are acousto-optic modulators (AOMs) driven by ultrasonic signals. These optical components (the selection optical elements AOM1-AOM6, the focusing lenses CD1-CD6, the reflectors M1-M12, the unit-side incident mirrors IM1-IM6, the collimating lenses CL1-CL6, and the absorber TR) are supported by a plate-shaped support unit IUB. This support unit IUB supports these optical components from below (in the -Z direction) above the multiple scanning units Un (U1-U6). Therefore, the supporting member IUB also has a function of thermally insulating the selective optical elements AOMn (AOM1 to AOM6) that serve as heat sources from the plurality of scanning units Un (U1 to U6).

The optical path of the light beam LB is bent into a zigzag shape by the reflectors M1 to M12 from the light source device 14', and the light beam LB is guided to the absorber TR. The following is a detailed description of the case where the selection optical elements AOMn (AOM1 to AOM6) are all in the Off state (a state where no ultrasonic signal is applied). The light beam LB (parallel light beam) from the light source device 14' travels in the +Y direction parallel to the Y axis, passes through the focusing lens CD1, and is incident on the reflector M1. The light beam LB reflected to the -X direction side by the reflector M1 is transmitted linearly from the first selection optical element AOM1 arranged at the focal position (beam waist position) of the focusing lens CD1, and becomes a parallel light beam again through the collimating lens CL1 and reaches the reflector M2. The light beam LB reflected to the +Y direction side by the reflector M2 passes through the focusing lens CD2 and is reflected to the +X direction side by the reflector M3.

The light beam LB reflected by the reflector M3 passes linearly through the second selective optical element AOM2, which is located at the focal position (beam waist position) of the condenser lens CD2. It then passes through the collimator lens CL2 and becomes a parallel beam again, reaching the reflector M4. The light beam LB reflected toward the +Y direction by the reflector M4 passes through the condenser lens CD3 and is then reflected toward the -X direction by the reflector M5. The light beam LB reflected toward the -X direction by the reflector M5 passes linearly through the third selective optical element AOM3, which is located at the focal position (beam waist position) of the condenser lens CD3. It then passes through the collimator lens CL3 and becomes a parallel beam again, reaching the reflector M6. The light beam LB reflected toward the +Y direction by the reflector M6 passes through the condenser lens CD4 and is then reflected toward the +X direction by the reflector M7.

The light beam LB reflected by the reflector M7 passes through the fourth selective optical element AOM4, which is located at the focal position (beam waist position) of the condenser lens CD4, in a straight line. It then passes through the collimator lens CL4 and becomes a parallel beam again, reaching the reflector M8. The light beam LB reflected toward the +Y direction by the reflector M8 passes through the condenser lens CD5 and is then reflected toward the -X direction by the reflector M9. The light beam LB reflected toward the -X direction by the reflector M9 passes through the fifth selective optical element AOM5, which is located at the focal position (beam waist position) of the condenser lens CD5, in a straight line. It then passes through the collimator lens CL5 and becomes a parallel beam again, reaching the reflector M10. The light beam LB reflected toward the +Y direction by the reflector M10 passes through the condenser lens CD6 and is then reflected toward the +X direction by the reflector M11. The light beam LB reflected by the reflector M11 passes linearly through the sixth selective optical element AOM6, which is located at the focal position (beam waist position) of the condenser lens CD6. It then passes through the collimator lens CL6 and becomes a parallel light beam again. After being reflected in the -Y direction by the reflector M12, it reaches the absorber TR. The absorber TR is a light absorber that absorbs the light beam LB to prevent it from leaking to the outside.

As described above, the selective optical elements AOM1-AOM6 are arranged so that the light beam LB from the light source device 14' is sequentially transmitted therethrough. Furthermore, the beam waists of the light beams LB are formed within each of the selective optical elements AOM1-AOM6 by the condenser lenses CD1-CD6 and the collimator lenses CL1-CL6. This reduces the diameter of the light beams LB incident on the selective optical elements AOM1-AOM6 (acousto-optic modulators), improving diffraction efficiency and responsiveness.

When an ultrasonic signal (high-frequency signal) is applied to each selection optical element AOMn (AOM1-AOM6), the incident light beam LB (zero-order light) is diffracted at a diffraction angle corresponding to the high-frequency frequency, generating first-order diffracted light as an outgoing light beam (light beam LBn). In the fourth embodiment, the light beams LBn emitted as first-order diffracted light from each of the plurality of selection optical elements AOMn (AOM1-AOM6) are designated as light beams LB1-LB6, and each selection optical element AOMn (AOM1-AOM6) is configured to deflect the optical path of the light beam LB from the light source device 14'. However, as described above, the actual efficiency of generating first-order diffracted light by an acousto-optic modulation element is approximately 80% of that of zero-order light. Therefore, the intensity of the light beams LB1-LB6 deflected by each selection optical element AOMn (AOM1-AOM6) is reduced compared to the original light beam LB. When any one of the selective optical elements AOMn (AOM1 to AOM6) is in the On state, approximately 20% of the zero-order light that travels straight without being diffracted remains, but is ultimately absorbed by the absorber TR.

Furthermore, the selective optical element AOMn is a diffraction grating that generates a periodic, coarse variation in refractive index in a predetermined direction within the transmissive member using ultrasonic waves. Therefore, when the incident light beam LB is linearly polarized (P-polarized or S-polarized), its polarization direction and the periodic direction of the diffraction grating are set so that the efficiency (diffraction efficiency) of generating first-order diffracted light is maximized. When the selective optical element AOMn is configured to diffract and deflect the incident light beam LB in the Z direction, as shown in FIG. 26 , the periodic direction of the diffraction grating generated within the selective optical element AOMn is also in the Z direction. Therefore, the polarization direction of the light beam LB from the light source device 14′ is set (adjusted) to match this direction.

As shown in FIG26 , the plurality of selection optical elements AOMn (AOM1 to AOM6) are each configured to deflect the deflected light beams LB1 to LB6 (first-order diffracted light) in the -Z direction relative to the incident light beam LB. The light beams LB1 to LB6 deflected and emitted from the selection optical elements AOMn (AOM1 to AOM6) are projected toward the unit-side incident mirrors IM1 to IM6, located at a predetermined distance from the selection optical elements AOMn (AOM1 to AOM6), and are thereby reflected in the -Z direction parallel to (coaxial with) the irradiation center axes Le1 to Le6. The light beams LB1 to LB6 reflected by the unit-side incident mirrors IM1 to IM6 (hereinafter simply referred to as mirrors IM1 to IM6) pass through openings TH1 to TH6 formed in the support member IUB, respectively, and are incident on the scanning units Un (U1 to U6) along the irradiation center axes Le1 to Le6.

The structures, functions, and effects of the various selection optical elements AOMn (AOM1 to AOM6) can be made identical to one another. The plurality of selection optical elements AOMn (AOM1 to AOM6) turns on/off the generation of diffracted light from the incident light beam LB, depending on the on/off state of the drive signal (high-frequency signal) from the control device 18. For example, when the selection optical element AOM1 is not applied with the drive signal (high-frequency signal) from the control device 18 and is in the off state, it does not diffract the incident light beam LB but transmits it. Therefore, the light beam LB that has passed through the selection optical element AOM1 is transmitted through the collimating lens CL1 and is incident on the reflective mirror M2. On the other hand, when the selection optical element AOM1 is applied with the drive signal from the control device 18 and is in the on state, it diffracts the incident light beam LB and directs it toward the mirror IM1. In other words, the selection optical element AOM1 is switched by the drive signal. The mirror IM1 reflects the light beam LB diffracted by the selection optical element AOM1 toward the scanning unit U1 side. The light beam LB1 reflected by the mirror IM1 passes through the opening TH1 of the support member IUB and is incident on the scanning unit U1 along the irradiation center axis Le1. Therefore, the mirror IM1 reflects the incident light beam LB1 so that the optical axis of the reflected light beam LB1 is coaxial with the irradiation center axis Le1. Furthermore, when the selection optical element AOM1 is in the on state, the zero-order light of the light beam LB that linearly passes through the selection optical element AOM1 (approximately 20% of the intensity of the incident light beam) is transmitted through the subsequent collimating lenses CL1 to CL6, focusing lenses CD2 to CD6, reflectors M2 to M12, and the selection optical elements AOM2 to AOM6, reaching the absorber TR.

FIG27A is a diagram showing the optical path switching of the light beam LB based on the selection optical element AOM1 as viewed from the +Z direction side, and FIG27B is a diagram showing the optical path switching of the light beam LB based on the selection optical element AOM1 as viewed from the -Y direction side. When the drive signal is in the Off state, the selection optical element AOM1 does not diffract the incident light beam LB but directly transmits it toward the side of the reflector M2. On the other hand, when the drive signal is in the On state, the selection optical element AOM1 generates a light beam LB1 that diffracts the incident light beam LB toward the -Z direction side and directs it toward the mirror IM1. Therefore, in the XY plane, the traveling directions of the light beam LB (zero-order light) emitted from the selection optical element AOM1 and the deflected light beam LB1 (first-order diffracted light) are not changed, and the traveling direction of the light beam LB1 (first-order diffracted light) is changed in the Z direction. In this way, the control device 18 switches the selection optical element AOM1 by turning the driving signal (high-frequency signal) to be applied to the selection optical element AOM1 On/Off (high/low), thereby switching whether the light beam LB is directed toward the subsequent selection optical element AOM2 or the deflected light beam LB1 is directed toward the scanning unit U1.

Similarly, when the drive signal (high-frequency signal) from the control device 18 is in the off state, the selective optical element AOM2 does not diffract the incident light beam LB (the light beam LB that is transmitted without being diffracted by the selective optical element AOM1) but transmits it toward the collimating lens CL2 side (the reflective mirror M4 side). When the drive signal from the control device 18 is in the on state, the diffracted light of the incident light beam LB, i.e., the light beam LB2, is directed toward the mirror IM2. The mirror IM2 reflects the light beam LB2 diffracted by the selective optical element AOM2 toward the scanning unit U2 side. The light beam LB2 reflected by the mirror IM2 passes through the opening TH2 of the support member IUB and is incident on the scanning unit U2 coaxially with the irradiation center axis Le2. Furthermore, when the drive signal (high-frequency signal) from the control device 18 is off, the selective optical elements AOM3-AOM6 do not diffract the incident light beam LB, but transmit it toward the collimating lenses CL3-CL6 (the reflective mirrors M6, M8, M10, and M12). When the drive signal from the control device 18 is on, the selective optical elements AOM3-AOM6 direct the primary diffracted light beams LB3-LB6 of the incident light beam LB toward the mirrors IM3-IM6. The mirrors IM3-IM6 reflect the light beams LB3-LB6 diffracted by the selective optical elements AOM3-AOM6 toward the scanning units U3-U6. The light beams LB3-LB6 reflected by the mirrors IM3-IM6 are coaxial with the irradiation center axes Le3-Le6, pass through the openings TH3-TH6 of the support member IUB, and enter the scanning units U3-U6. In this way, the control device 18 switches one of the selection optical elements AOM2 to AOM6 by turning the driving signal (high-frequency signal) to be applied to each of the selection optical elements AOM2 to AOM6 On/Off (high/low), thereby switching whether the light beam LB is directed toward the subsequent selection optical element AOM3 to AOM6 or the absorber TR, or toward the corresponding scanning unit U2 to U6.

As described above, the light beam switching component 20 includes a plurality of selection optical elements AOMn (AOM1 to AOM6) arranged in series along the direction of travel of the light beam LB from the light source device 14'. This allows the optical path of the light beam LB to be switched to select a scanning unit Un for the light beam LBn to be incident on. For example, to cause the light beam LB1 to be incident on the scanning unit U1, the selection optical element AOM1 can be turned on. To cause the light beam LB3 to be incident on the scanning unit U3, the selection optical element AOM3 can be turned on. The plurality of selection optical elements AOMn (AOM1 to AOM6) are provided corresponding to the plurality of scanning units Un (U1 to U6) to switch whether the light beam LBn is incident on the corresponding scanning unit Un.

Because multiple scanning units Un (U1-U6) repeatedly scan the spot light SP in a predetermined order, the beam switching component 20 switches the scanning unit U1-U6 on which one of the light beams LB1-LB6 is incident. For example, if the order of scanning units Un scanning the spot light SP is U1 → U2 → ... → U6, the beam switching component 20 switches the scanning unit Un on which the light beam LBn is incident in the order U1 → U2 → ... → U6.

According to the above, each selection optical element AOMn (AOM1 to AOM6) of the light beam switching component 20 only needs to be in the On state during one scan of the point light SP based on the polygon mirror PM of the scanning unit Un (U1 to U6). The details will be described later. If the number of reflection surfaces of the polygon mirror PM is set to Np and the rotation speed of the polygon mirror PM is set to Vp (rpm), the time Tss corresponding to the rotation angle of one side of the reflection surface RP of the polygon mirror PM becomes Tss = 60/(Np·Vp) (seconds). For example, when the number of reflection surfaces Np is 8 and the rotation speed Vp is 30,000, one rotation of the polygon mirror PM is 2 milliseconds and the time Tss is 0.25 milliseconds. Converting this into a frequency is 4 kHz, which means that compared with the acousto-optic modulator used to modulate the light beam LB with a wavelength in the ultraviolet region at a high speed of several tens of MHz in response to the drawing data, it can be an acousto-optic modulator with a relatively low response frequency. Therefore, it is possible to use an acousto-optic modulation element with a large diffraction angle for the light beams LB1 to LB6 (first-order diffracted light) deflected relative to the incident light beam LB (zero-order light), and it becomes easy to configure the mirrors IM1 to IM6 (Figures 26, 27A, and 27B) that separate the light beams LB1 to LB6 deflected relative to the travel path of the light beam LB passing straight through the selected optical elements AOM1 to AOM6.

Furthermore, the multiple scanning units U1-U6 repeatedly scan the spot light SP once in a predetermined sequence. Accordingly, the serial data DLn of the pattern data of each scanning unit Un is output to the driver circuit 206a of the light source device 14' in a predetermined sequence. The serial data DLn sequentially output to the driver circuit 206a is referred to as the drawing bit string data Sdw. For example, if the predetermined sequence is U1 → U2 → ... → U6, first, one column of serial data DL1 is output to the driver circuit 206a, followed by one column of serial data DL2. In this manner, one column of serial data DL1-DL6, constituting the drawing bit string data Sdw, is sequentially output to the driver circuit 206a. Then, the next column of serial data DL1-DL6 is sequentially output to the driver circuit 206a as the drawing bit string data Sdw. The specific structure for outputting the drawing bit string data Sdw to the driver circuit 206a will be described in detail later.

The structure of the scanning unit Un (U1 to U6) can be the structure used in the first to third embodiments described above, but in this fourth embodiment, a scanning unit Un having the structure shown in Figure 28 is used. In addition, the scanning unit Un described below can also be used as the scanning unit of the first to third embodiments described above.

Hereinafter, the optical structure of the scanning unit Un (U1 to U6) used in the fourth embodiment will be described with reference to FIG28. In addition, since each scanning unit Un (U1 to U6) has the same structure, only the scanning unit U1 will be described, and the description of the other scanning units Un will be omitted. In addition, in FIG28, the direction parallel to the irradiation center axis Len (Le1) is set as the Zt direction, the direction on the plane perpendicular to the Zt direction and from the processing device PR1 to the exposure device EX toward the processing device PR2 of the substrate FS is set as the Xt direction, and the direction on the plane perpendicular to the Zt direction and perpendicular to the Xt direction is set as the Yt direction. That is, the three-dimensional coordinates of Xt, Yt, and Zt in FIG28 are three-dimensional coordinates obtained by rotating the three-dimensional coordinates of X, Y, and Z in FIG23 with the Y axis as the center so that the Z axis direction is parallel to the irradiation center axis Len (Le1).

As shown in FIG28 , within the scanning unit U1, along the direction of travel of the light beam LB1 from its incident position to the irradiated surface of the substrate FS, are provided a reflector M20, a beam expander BE, a reflector M21, a polarizing beam splitter BS, a reflector M22, an image-shifting optical component SR, a field stop FA, a reflector M23, a λ/4 wave plate QW, a cylindrical lens CYa, a reflector M24, a polygonal mirror PM, an fθ lens FT, a reflector M25, and a cylindrical lens CYb. Furthermore, within the scanning unit U1, there are provided an optical lens system G10 and a photodetector DTS for detecting light reflected from the irradiated surface of the substrate FS via the polarizing beam splitter BS.

The light beam LB1 incident on the scanning unit U1 travels in the -Zt direction and enters the reflector M20, which is tilted 45° relative to the XtYt plane. The axis of the light beam LB1 incident on the scanning unit U1 is coaxial with the irradiation center axis Le1. The reflector M20 functions as an incident optical component for causing the light beam LB1 to enter the scanning unit U1, reflecting the incident light beam LB1 in the -Xt direction toward the reflector M21 along an optical axis set parallel to the Xt axis. Therefore, the optical axis of the light beam LB1 traveling parallel to the Xt axis is orthogonal to the irradiation center axis Le1 within a plane parallel to the XtZt plane. The light beam LB1 reflected by the reflector M20 passes through the beam expander BE, which is arranged along the optical axis of the light beam LB1 traveling parallel to the Xt axis, and enters the reflector M21. The beam expander BE expands the diameter of the transmitted light beam LB1. The beam expander BE includes a condenser lens Be1 and a collimator lens Be2 for converting the light beam LB1 , which is converged and then diverged by the condenser lens Be1 , into parallel light.

The reflector M21 is tilted 45° relative to the YtZt plane and reflects the incident light beam LB1 toward the polarizing beam splitter BS in the -Yt direction. The polarization separation plane of the polarizing beam splitter BS is tilted 45° relative to the YtZt plane, reflecting the P-polarized light beam while transmitting a linearly polarized (S-polarized) light beam polarized in a direction orthogonal to the P-polarized light beam. The light beam LB1 incident on the scanning unit U1 is P-polarized. Therefore, the polarizing beam splitter BS reflects the light beam LB1 from the reflector M21 in the -Xt direction and directs it toward the reflector M22.

Mirror M22 is arranged at a 45° angle relative to the XtYt plane and reflects the incident light beam LB1 in the -Zt direction toward mirror M23, which is spaced apart from mirror M22 in the -Zt direction. Light beam LB1 reflected by mirror M22 passes through the image-shifting optical component SR and the field stop (field diaphragm) FA along an optical axis parallel to the Zt axis and is incident on mirror M23. The image-shifting optical component SR two-dimensionally adjusts the center position of the cross section of light beam LB1 within a plane (XtYt plane) perpendicular to the direction of travel of light beam LB1. The image-shifting optical component SR consists of two quartz parallel plates Sr1 and Sr2, arranged along the optical axis of light beam LB1 traveling parallel to the Zt axis. Parallel plate Sr1 can tilt about the Xt axis, and parallel plate Sr2 can tilt about the Yt axis. By tilting the parallel plates Sr1 and Sr2 about the Xt and Yt axes, respectively, the center position of the light beam LB1 is slightly displaced two-dimensionally in the XtYt plane perpendicular to the direction of travel of the light beam LB1. The parallel plates Sr1 and Sr2 are driven by an actuator (drive unit) (not shown) under the control of the control device 18.

Light beam LB1, having passed through the image-shifting optical component SR, passes through the circular opening of the field stop FA and reaches the reflector M23. The circular opening of the field stop FA blocks the peripheral portion of the intensity distribution within the cross section of light beam LB1, expanded by the beam expander BE. If the aperture of the circular opening of the field stop FA is configured as an adjustable variable iris diaphragm, the intensity (brightness) of the spot light SP can be adjusted.

The reflector M23 is arranged at an angle of 45° relative to the XtYt plane, and reflects the incident light beam LB1 in the +Xt direction toward the reflector M24 separated from the reflector M23 in the +Xt direction. The light beam LB1 reflected by the reflector M23 is transmitted through the λ/4 wave plate QW and the cylindrical lens CYa and is incident on the reflector M24. The reflector M24 reflects the incident light beam LB1 toward the polygon mirror (rotating polygon mirror, scanning deflection component) PM. The polygon mirror PM reflects the incident light beam LB1 in the +Xt direction toward the fθ lens FT having an optical axis AXf parallel to the Xt axis. The polygon mirror PM deflects (reflects) the incident light beam LB1 in a plane parallel to the XtYt plane in order to scan the point light SP of the light beam LB1 on the irradiated surface of the substrate FS. Specifically, the polygon mirror PM has a rotation axis AXp extending in the Zt axis direction, and a plurality of reflection surfaces RP (eight reflection surfaces RP in this fourth embodiment) formed around the rotation axis AXp. By rotating the polygonal mirror PM in a predetermined direction about the rotation axis AXp, the reflection angle of the pulsed light beam LB1 irradiated on the reflection surface RP can be continuously changed. Thus, by deflecting the reflection direction of the light beam LB1 by the reflection surface RP, the point light SP of the light beam LB1 irradiated on the irradiated surface of the substrate FS can be scanned along the scanning direction (the width direction of the substrate FS, the Yt direction).

The point light SP of the light beam LB1 can be scanned along the drawing line SL1 through a reflecting surface RP. Therefore, with one rotation of the polygonal mirror PM, the maximum number of drawing lines SL1 scanned by the point light SP on the irradiated surface of the substrate FS is eight, which is the same as the number of reflecting surfaces RP. The polygonal mirror PM rotates at a certain speed by the polygonal mirror driving unit RM including a motor, etc. The rotation of the polygonal mirror PM based on the polygonal mirror driving unit RM is controlled by the control device 18. As described above, the effective length of the drawing line SL1 (for example, 30 mm) is set to a length less than the maximum scanning length (for example, 31 mm) that can be used to scan the point light SP through the polygonal mirror PM. In the initial setting (design), the center point of the drawing line SL1 (the point through which the irradiation center axis Le1 passes) is set in the center of the maximum scanning length.

Furthermore, as an example, if the effective length of the drawing line SL1 is set to 30 mm, and the spot light SP with an effective size of 3 μm is irradiated onto the irradiated surface of the substrate FS along the drawing line SL1 while overlapping each other by 1.5 μm, the number of spot light SP irradiated in one scan (the number of pulses of the light beam LB from the light source device 14') is 20,000 (30 mm/1.5 μm). Furthermore, if the scanning time of the spot light SP along the drawing line SL1 is set to 200 μsec, the pulsed spot light SP must be irradiated 20,000 times during this period. Therefore, the light emission frequency Fs of the light source device 14' becomes Fs ≥ 20,000 times/200 μsec = 100 MHz.

In the non-scanning direction (Zt direction) orthogonal to the scanning direction (rotational direction) achieved by the polygon mirror PM, the cylindrical lens CYa converges the incident light beam LB1 into a slit shape on the reflection surface RP of the polygon mirror PM. The cylindrical lens CYa, whose generatrix is parallel to the Yt direction, suppresses the influence of the reflection surface RP being tilted relative to the Zt direction (the reflection surface RP being tilted relative to the normal of the XtYt plane), thereby preventing the irradiation position of the light beam LB1 on the irradiated surface of the substrate FS from being shifted in the Xt direction.

The fθ lens FT, having an optical axis AXf extending along the Xt axis, is a telecentric scanning lens that projects the light beam LB1 reflected by the polygon mirror PM toward the reflector M25 in the XtYt plane, parallel to the optical axis AXf. The incident angle θ of the light beam LB1 on the fθ lens FT varies depending on the rotation angle (θ/2) of the polygon mirror PM. The fθ lens FT projects the light beam LB1, via the reflector M25 and cylindrical lens CYb, to an image height position on the illuminated surface of the substrate FS that is proportional to the incident angle θ. If the focal length is fo and the image height position is y, the fθ lens FT is designed to satisfy the relationship y = fo·θ. Therefore, the fθ lens FT can accurately and uniformly scan the light beam LB1 (point light SP) in the Yt direction (Y direction). When the incident angle θ on the fθ lens FT is 0 degrees, the light beam LB1 incident on the fθ lens FT travels along the optical axis AXf.

The reflector M25 reflects the incident light beam LB1 toward the substrate FS along the -Zt direction via the cylindrical lens CYb. The light beam LB1 projected onto the substrate FS is converged into a tiny point light SP with a diameter of several μm (for example, 3 μm) on the irradiated surface of the substrate FS by the fθ lens FT and the cylindrical lens CYb whose main line is parallel to the Yt direction. In addition, the point light SP projected onto the irradiated surface of the substrate FS is scanned in one dimension along the drawing line SL1 extending in the Yt direction by the polygon mirror PM. In addition, the optical axis AXf of the fθ lens FT and the irradiation center axis Le1 are located on the same plane, which is parallel to the XtZt plane. Therefore, the light beam LB1 traveling on the optical axis AXf is reflected in the -Zt direction by the reflector M25 and is projected onto the substrate FS coaxially with the irradiation center axis Le1. In this fourth embodiment, at least the fθ lens FT functions as a projection optical system that projects the light beam LB1 deflected by the polygon mirror PM onto the irradiated surface of the substrate FS. Furthermore, at least the reflective components (mirrors M21-M25) and polarizing beam splitter BS function as optical path deflection components that bend the optical path of light beam LB1 from reflector M20 to substrate FS. This optical path deflection component enables the incident axis of light beam LB1 incident on reflector M20 to be approximately coaxial with the irradiation center axis Le1. On the XtZt plane, light beam LB1 passing through scanning unit U1 follows a roughly U-shaped or C-shaped optical path before traveling in the -Zt direction and projecting onto substrate FS.

In this manner, while the substrate FS is being transported in the X direction, each scanning unit Un (U1 to U6) causes the point light SP of the light beam LBn to perform one-dimensional scanning in the scanning direction (Y direction). This allows the point light SP to perform two-dimensional scanning relative to the irradiated surface of the substrate FS. Consequently, a predetermined pattern can be drawn and exposed in the exposure area W of the substrate FS.

The light detector DTS has a photoelectric conversion element that performs photoelectric conversion on the incident light. A predetermined reference pattern is formed on the surface of the rotating cylinder DR. The portion of the rotating cylinder DR on which the reference pattern is formed is made of a material having a low reflectivity (10 to 50%) with respect to the wavelength band of the light beam LB1, and the other portions of the rotating cylinder DR on which the reference pattern is not formed are made of a material having a reflectivity of less than 10% or a material that absorbs light. Therefore, if the point light SP of the light beam LB1 is irradiated from the scanning unit U1 to the area of the rotating cylinder DR on which the reference pattern is formed in a state where the substrate FS is not wound (or in a state where it passes through the transparent portion of the substrate FS), the reflected light passes through the cylindrical lens CYb, the reflector M25, the fθ lens FT, the polygonal mirror PM, the reflector M24, the cylindrical lens CYa, the λ/4 wave plate QW, the reflector M23, the field aperture FA, the image-shifting optical component SR and the reflector M22 and is incident on the polarization beam splitter BS. Here, a λ/4 wave plate QW is provided between the polarization beam splitter BS and the substrate FS, specifically, between the reflector M23 and the cylindrical lens CYa. Thus, the light beam LB1 incident on the substrate FS is converted from P-polarized light to circularly polarized light by the λ/4 wave plate QW. Meanwhile, the light reflected from the substrate FS incident on the polarization beam splitter BS is converted from circularly polarized light by the λ/4 wave plate QW. Consequently, the light reflected from the substrate FS passes through the polarization beam splitter BS and, via the optical lens system G10, is incident on the photodetector DTS.

At this time, in a state where the pulsed light beam LB1 (preferably the light beam LB1 from the seed light S1) is continuously incident on the scanning unit U1, the rotating cylinder DR rotates and the scanning unit U1 causes the point light SP to scan, thereby irradiating the outer peripheral surface of the rotating cylinder DR in two dimensions. Therefore, the image of the reference pattern formed on the rotating cylinder DR can be obtained by the light detector DTS. Specifically, the intensity change of the photoelectric signal output from the light detector DTS is made in response to the clock pulse signal for the pulsed emission of the point light SP (generated in the light source device 14'), and digital sampling is performed at each scanning time, thereby obtaining one-dimensional image data in the Yt direction. Moreover, in response to the measurement value of the encoder ENn that measures the rotation angle position of the rotating cylinder DR, the one-dimensional image data in the Yt direction is arranged in the Xt direction at a certain distance in the sub-scanning direction (for example, 1/2 of the size of the point light SP), thereby obtaining two-dimensional image information of the surface of the rotating cylinder DR. Based on the acquired two-dimensional image information of the reference pattern on the rotating drum DR, the control device 18 measures the inclination of the drawing line SL1 of the scanning unit U1. The inclination of the drawing line SL1 can be the relative inclination between the scanning units Un (U1-U6) or the inclination relative to the central axis Axo of the rotating drum DR (absolute inclination). Similarly, the inclinations of the drawing lines SL2-SL6 can also be measured.

As shown in FIG29 , an origin sensor (origin detector) OP1 is provided around the polygon mirror PM of the scanning unit U1. The origin sensor OP1 outputs a pulsed origin signal SZ indicating the start of scanning of the point light SP based on each reflecting surface RP. The origin sensor OP1 outputs the origin signal SZ when the rotation position of the polygon mirror PM reaches a predetermined position immediately before the start of scanning of the point light SP based on the reflecting surface RP. Since the polygon mirror PM can deflect the light beam LB1 projected onto the substrate FS within the scanning angle range θs, if the reflection direction (deflection direction) of the light beam LB1 reflected by the polygon mirror PM is within the scanning angle range θs, the reflected light beam LB1 is incident on the fθ lens FT. Therefore, the origin sensor OP1 outputs the origin signal SZ when the rotation position of the polygon mirror PM reaches a predetermined position immediately before the reflection direction of the light beam LB1 reflected by the reflecting surface RP enters the scanning angle range θs. In addition, the scanning angle range θs and the maximum scanning rotation angle range α shown in FIG7 have a relationship of θs=2×α.

Since the polygon mirror PM has eight reflecting surfaces RP, the origin sensor OP1 outputs an origin signal SZ eight times during one rotation of the polygon mirror PM. The origin signal SZ detected by the origin sensor OP1 is transmitted to the control device 18. After the origin sensor OP1 outputs the origin signal SZ, the spot light SP begins scanning along the drawing line SL1.

The origin sensor OP1 outputs an origin signal SZ using a reflection surface RP adjacent to the reflection surface RP on which the spot light SP is subsequently scanned (deflected by the light beam LB1) (in the fourth embodiment, the reflection surface RP immediately preceding the polygon mirror PM in the rotational direction). To facilitate differentiation between the reflection surfaces RP, the reflection surface RP currently deflecting the light beam LB1 is designated RPa in FIG. 29 , while the other reflection surfaces RP located counterclockwise (in the direction opposite to the rotational direction of the polygon mirror PM) are designated RPb to RPh.

The origin sensor OP1 includes a beam transmitting system Opa comprising a light source 312 that emits a laser beam Bga in a non-photosensitive wavelength range, such as a semiconductor laser, and mirrors 314 and 316 that reflect the laser beam Bga from the light source 312 and project it onto the reflecting surface RPb of the polygon mirror PM. Furthermore, the origin sensor OP1 includes a beam receiving system Opb comprising a light receiving unit 318, mirrors 320 and 322 that guide the reflected light (reflected beam Bgb) of the laser beam Bga reflected by the reflecting surface RPb toward the light receiving unit 318, and a lens system 324 that focuses the reflected beam Bgb reflected by the mirror 322 into a tiny spot of light. The light receiving unit 318 includes a photoelectric conversion element that converts the spot of reflected beam Bgb focused by the lens system 324 into an electrical signal. Here, the position where the laser beam Bga is projected onto each reflecting surface RP of the polygon mirror PM is set to the pupil plane (the position of the focal point) of the lens system 324 .

The beam sending system Opa and the beam receiving system Opb are located at positions where the beam receiving system Opb can receive the reflected beam Bgb of the laser beam Bga emitted by the beam sending system Opa when the rotational position of the polygon mirror PM reaches a predetermined position immediately before scanning of the point light SP by the reflection surface RP begins. In other words, the beam sending system Opa and the beam receiving system Opb are located at positions where they can receive the reflected beam Bgb of the laser beam Bga emitted by the beam sending system Opa when the angle of the reflection surface RP reaches a predetermined angular position. Furthermore, reference numeral Msf in FIG29 denotes the shaft of the rotary motor of the polygon mirror driving unit RM (see FIG28 ), which is arranged coaxially with the rotation axis AXp.

A light shield (not shown) with a narrow slit opening is located immediately before the light-receiving surface of the photoelectric conversion element within the light-receiving section 318. While the angular position of the reflecting surface RPb is within a predetermined angular range, the reflected light beam Bgb is incident on the lens system 324. The point light of the reflected light beam Bgb scans along a predetermined direction on the light shield within the light-receiving section 318. During this scanning process, the point light of the reflected light beam Bgb that passes through the slit opening of the light shield is received by the photoelectric conversion element within the light-receiving section 318. The resulting light signal is amplified by an amplifier and output as a pulsed origin signal SZ.

As described above, the origin sensor OP1 detects the origin signal SZ using the preceding reflection surface RPb in the rotational direction, using the reflection surface RPa that deflects the light beam LB1 (scans the spot light SP). Therefore, if the angle ηj between adjacent reflection surfaces RP (for example, reflection surface RPa and reflection surface RPb) differs from the designed value (135 degrees when there are eight reflection surfaces RP), the timing of generating the origin signal SZ may differ for each reflection surface RP due to this deviation, as shown in FIG30 .

In FIG30 , the origin signal SZ generated using the reflection surface RPb is set to SZb. Similarly, the origin signals SZ generated using the reflection surfaces RPc, RPd, RPe, ... are set to SZc, SZd, SZe, .... When the angle ηj formed by the adjacent reflection surfaces RP of the polygon mirror PM is the design value, the interval between the generation timings of the origin signals SZ (SZb, SZc, SZd, ...) is the time Tpx. The prescribed time Tpx is the time required for the polygon mirror PM to rotate one side of the reflection surface RP. However, in FIG30 , due to the error in the angle ηj formed by the reflection surfaces RP of the polygon mirror PM, the timing of the origin signals SZc and SZd generated using the reflection surfaces RPc and RPd is offset from the regular generation timing. In addition, the time intervals Tp1, Tp2, Tp3, ... generated by the origin signals SZb, SZc, SZd, SZe, ... are not constant at the microsecond level due to manufacturing errors in the polygon mirror PM. In the time diagram shown in Figure 30, Tp1 < Tpx, Tp2 > Tpx, and Tp3 < Tpx. Furthermore, if the number of reflecting surfaces RP is Np and the rotation speed of the polygon mirror PM is Vp, Tpx is expressed as Tpx = 60/(Np × Vp) (seconds). For example, if Vp is 30,000 rpm and Np is 8, Tpx becomes 250 μ seconds.

Therefore, due to the error in the angle ηj formed by the adjacent reflecting surfaces RP of the polygonal mirror PM, the position of the drawing starting point (scanning starting point) of the drawing line SL1 drawn by each reflecting surface RP (RPa to RPh) on the irradiated surface of the substrate FS will be offset along the main scanning direction. As a result, the position of the drawing end point of the drawing line SL1 will also be offset along the main scanning direction. In other words, the position of the drawing line SL1 of the point light SP drawn by each reflecting surface RP is displaced along the scanning direction (Y direction), so the positions of the drawing starting point and the drawing end point of each drawing line SLn will not form a straight line along the X direction. The main reason for the deviation of the positions of the drawing starting point and the drawing end point of the drawing line SL1 of the point light SP along the main scanning direction is that Tp1, Tp2, Tp3,...=Tpx does not hold.

Therefore, in this fourth embodiment, as shown in the timing chart of Figure 30 , the drawing start point of the spot light SP is determined to be the time Tpx elapsed after the generation of a pulsed origin signal SZ. Specifically, after the time Tpx elapses after the generation of the origin signal SZ, the control device 18 controls the beam switching component 20 so that the light beam LB1 is incident on the scanning unit U1. Furthermore, the control device 18 outputs the drawing bit string data Sdw, or serial data DL1, for the scanning unit U1 to be scanned next, to the driver circuit 206a of the light source device 14' shown in Figure 26 . This allows the reflection surface RPb used to detect the origin signal SZ and the reflection surface RP used to actually scan the spot light SP to be the same reflection surface.

Specifically, after the origin sensor OP1 of the scanning unit U1 outputs the origin signal SZb, the control device 18 outputs an On drive signal to the selection optical element AOM1 of the beam switching component 20 for a certain period of time (On time Ton) after a time Tpx has passed. The certain period of time (On time Ton) during which the selection optical element AOM1 is On is a predetermined period of time, and is set to cover the period (scanning period) during which the point light SP is scanned once along the drawing line SL1 by one of the reflection surfaces RP of the polygon mirror PM. Next, the control device 18 outputs the serial data DL1 of a specific column, for example, the first column, to the drive circuit 206a of the light source device 14'. As a result, during the scanning period during which the scanning unit U1 scans the point light SP, the light beam LB1 is incident on the scanning unit U1, so that the scanning unit U1 can draw a pattern corresponding to the serial data DL1 of the specific column (for example, the first column). As described above, since the scanning unit U1 scans the spot light SP after a time Tpx has passed since the origin sensor OP1 of the scanning unit U1 outputs the origin signal SZb, the reflecting surface RPb for detecting the origin signal SZb can scan the spot light SP caused by the origin signal SZb.

Next, after a time Tpx has passed since the origin sensor OP1 of the scanning unit U1 outputted the origin signal SZd, the control device 18 outputs an On drive signal to the selection optical element AOM1 of the light beam switching component 20 for a predetermined time (On time Ton). The control device 18 then outputs the next column of serial data DL1, for example, the second column, to the drive circuit 206a of the light source device 14'. Thus, since the light beam LB1 is incident on the scanning unit U1 during a period that includes the time required for the scanning unit U1 to scan the spot light SP, the scanning unit U1 can draw a pattern corresponding to the next column of serial data DL1 (for example, the second column). Thus, since the scanning unit U1 scans the spot light SP after a time Tpx has passed since the origin sensor OP1 of the scanning unit U1 outputted the origin signal SZd, the spot light SP caused by the origin signal SZd can be scanned by the reflection surface RPb used to detect the origin signal SZd. In addition, when scanning the spot light SP is performed not on each continuous reflection surface RP of the polygon mirror PM but on a skipped surface, the origin signal SZ is used to perform drawing processing in a skipped manner. The reason for skipping one will be described in detail later.

As described above, after time Tpx has passed since the origin sensor OP1 of the scanning unit U1 outputs the origin signal SZ, the control device 18 controls the light beam switching component 20 so that the scanning unit U1 causes the spot light SP to scan, and also outputs serial data DL1 to the drive circuit 206a of the light source device 14'. Furthermore, each time scanning by the scanning unit U1 begins, the control device 18 shifts the columns of the output serial data DL1 in the column direction, such as the first column, the second column, the third column, the fourth column, and so on. Furthermore, between one scan of the spot light SP by the scanning unit U1 and the next scan, the other scanning units Un (scanning units U2 to U6) sequentially scan the spot light SP. The scanning of the spot light SP by the other scanning units Un (U2 to U6) is the same as that by the scanning unit U1. Furthermore, an origin sensor OPn (OP1 to OP6) is provided for each scanning unit Un (U1 to U6).

As described above, by using the reflecting surface RP for detecting the origin signal SZb of the scanning unit U1 to perform scanning of the point light SP, even if there is an error in the angle ηj between the adjacent reflecting surfaces RP of the polygonal mirror PM, the positions of the starting point and the ending point of the drawing of the point light SP drawn by each reflecting surface RP (RPa~RPh) on the irradiated surface of the substrate FS can be prevented from being offset along the main scanning direction.

Therefore, the time Tpx for the polygon mirror PM to rotate 45 degrees needs to be accurate at the μ-second level, that is, the speed of the polygon mirror PM needs to be uniform and precisely rotated at a constant speed. When the polygon mirror PM is rotated at a constant speed in this way, the reflection surface RP used to generate the origin signal SZ becomes an angle that accurately rotates only 45 degrees after the time Tpx to reflect the light beam LB1 toward the fθ lens FT. Therefore, by improving the rotational uniformity of the polygon mirror PM and minimizing the speed unevenness in one rotation, the position of the reflection surface RP used to generate the origin signal SZ can be made different from the position of the reflection surface RP used to deflect the light beam LB1 and scan the point light SP. In other words, since the generation timing of the origin signal SZ is delayed by the time Tpx, the result is the same effect as when the origin signal SZ is detected using the reflection surface RP that scans the point light SP. As a result, the configuration freedom of the origin sensor OP1 (OPn) is improved, and an origin sensor with high rigidity and stable structure can be set. Furthermore, although the reflective surface RP that origin sensor OP1 (OPn) detects is the one preceding in the rotational direction of the reflective surface RP that deflects light beam LB1 (LBn), it is not limited to the preceding one as long as it is located before the polygon mirror PM in the rotational direction. In this case, if the reflective surface RP that origin sensor OP detects is the first n (an integer greater than or equal to 1) reflective surfaces RP in the rotational direction of the reflective surface RP that deflects light beam LB1 (LBn), the drawing start point can be set after n × time Tpx has passed since the origin signal SZ was generated.

Furthermore, for every other origin signal SZb, SZd, etc. generated by origin sensor OP1 (OPn), the drawing start point is set n times Tpx later. This provides a margin in the processing time for reading the pixel data sequence corresponding to each drawing line SL1, transmitting (communicating) data, performing correction calculations, and so on. Consequently, pixel data sequence transmission errors, errors in pixel data sequences, and/or partial loss can be reliably avoided.

Furthermore, instead of providing an origin sensor OPn for detecting a reflection surface RP adjacent to the reflection surface RP on which the next point light SP is scanned (deflection of the light beam LB1) (in the fourth embodiment, the previous reflection surface RP in the rotational direction of the polygon mirror PM) as shown in FIG29 above, an origin sensor for detecting the same reflection surface RP as the reflection surface RP on which the next point light SP is scanned (deflection of the light beam LB1) may be provided. In this case, as illustrated in FIG30 , since the time intervals of the origin signals (pulse-like) SZ generated by the respective reflection surfaces RPa to RPh of the polygon mirror PM deviate, it is necessary to add a time offset corresponding to the amount of the deviation for each reflection surface RPa to RPh.

As also illustrated in FIG7 , when the number Np of the reflective surfaces RP of the polygon mirror PM is eight and the maximum scanning rotation angle range α is 15 degrees, the scanning efficiency (α/β) becomes 1/3. For example, between the time when scanning unit U1 scans the spot light SP and the time it performs the next scan, the light beam LBn can be distributed to two scanning units Un other than scanning unit U1 to scan the spot light SP. In other words, during the time when the polygon mirror PM of scanning unit U1 rotates one plane, a corresponding light beam LBn can be distributed to each of the three scanning units Un, including scanning unit U1, to scan the spot light SP.

However, since the scanning efficiency of the polygonal mirror PM is 1/3, when each scanning unit Un scans the point light SP with a maximum scanning rotation angle range α (15 degrees), the light beam LBn cannot be distributed to more than three scanning units Un (U2 to U6) other than the scanning unit U1 during the period when the polygonal mirror PM of the scanning unit U1 rotates one side of the reflection surface RP (β = 45 degrees). In other words, from the start of scanning of the point light SP of the scanning unit U1 to the start of the next scanning of the point light SP, the light beam LBn cannot be distributed to more than three scanning units Un (U2 to U6) other than the scanning unit U1. Therefore, in order to distribute the light beam LBn to each of the other five scanning units Un (U2 to U6) to perform scanning based on the point light SP during the period from the start of scanning of the point light SP of the scanning unit U1 to the start of the next scanning, the following method can be considered.

Even when the maximum scanning rotation angle range α is 15 degrees, the scanning rotation angle range α' of the polygonal mirror PM that can actually scan the point light SP is set to be smaller than the maximum scanning rotation angle range α (α = 15 degrees). Specifically, during the period when the polygonal mirror PM of each scanning unit Un (U1 to U6) rotates by the amount of one side of the reflecting surface RP (β = 45 degrees), the number of scanning units Un to which the light beam LBn is to be distributed is six, so the scanning rotation angle range α' is set to α' = 45/6 = 7.5 degrees. That is, the oscillation angle centered on the optical axis AXf of the light beam LBn incident on the fθ lens FT in Figure 28 is limited to ±7.5 degrees. Thus, during the period when the polygonal mirror PM of each scanning unit Un rotates 45 degrees (the period when the reflecting surface RP rotates by the amount of one side), the light beam LBn can be sequentially distributed and incident on one of the six scanning units Un (U1 to U6), and the scanning units Un (U1 to U6) can sequentially perform scanning based on the point light SP. However, if this is the case, the actual scanning rotation angle range α' that the spot light SP can scan becomes too small, and the maximum scanning range length of the spot light SP scan, that is, the maximum scanning length of the traced line SLn, becomes too short. To avoid this problem, a long focal length fθ lens FT is prepared without changing the maximum scanning length of the spot light SP scan, and the distance (working distance) from the reflection surface RP of the polygon mirror PM to the fθ lens FT is set long. In this case, there is a concern that the fθ lens FT will be larger, the Xt-direction dimensions of the scanning units Un (U1-U6) will be larger, and the stability of the beam scanning will be reduced due to the long working distance.

On the other hand, considering reducing the number of reflection surfaces RP of the polygonal mirror PM, the rotation angle β of the amount of one side of the reflection surface RP rotated by the polygonal mirror PM is increased. In this case, the shortening of the drawing line SLn and the enlargement of the scanning unit Un (U1 to U6) are suppressed, and at the same time, during the period when the polygonal mirror PM of the scanning unit Un (U1 to U6) rotates the reflection surface RP by the amount of one side (rotation angle β), the light beam LBn can be distributed and the six scanning units Un (U1 to U6) can sequentially scan the point light SP. For example, when the number of reflection surfaces RP of the polygonal mirror PM is set to four, that is, when the shape of the polygonal mirror PM is set to a square, the rotation angle β of the amount of one side of the reflection surface RP of the polygonal mirror PM becomes 90 degrees. Therefore, when the light beam LBn is distributed and the point light SP is scanned by six scanning units Un (U1~U6) during the period when the polygonal mirror PM of the scanning unit U1 rotates one side of the reflecting surface RP, the scanning rotation angle range α' of the polygonal mirror PM that can actually scan the point light SP becomes α'=90/6=15 degrees, which is equal to the above-mentioned maximum scanning rotation angle range α.

However, if a polygonal mirror PM with a small number of reflecting surfaces Np, such as a triangle or square, is rotated at high speed, the air resistance (wind resistance) becomes too large, and the rotation speed and the number of revolutions decrease (regularity). For example, even if the polygonal mirror PM is intended to rotate at a high speed of tens of thousands of rpm (rotations per minute), the rotation speed will be reduced by about 20% to 30% due to air resistance, and the desired high rotation speed and high number of revolutions cannot be obtained. In addition, a method of increasing the size of the polygonal mirror PM is also considered, but the weight of the polygonal mirror PM becomes too large, and the desired high rotation speed and high number of revolutions cannot be obtained. In addition, as a method of reducing the wind resistance during rotation even if the number of reflecting surfaces Np of the polygonal mirror PM is reduced, it is also considered to place the entire polygonal mirror PM in a vacuum environment, or in an environment of a gas with a smaller molecular weight than air (helium, etc.). In this case, an airtight structure for creating such an environment is provided around the polygonal mirror PM, which will correspondingly lead to an increase in the size of the scanning unit Un (U1 to U6).

Therefore, in the fourth embodiment, a polygonal mirror PM having a relatively large number of reflecting surfaces Np, that is, an octagonal shape that is closer to a circle, is used, and the scanning rotation angle range α' of the polygonal mirror PM that can actually scan the point light SP is set to the maximum scanning rotation angle range α (α = 15 degrees), and the reflecting surface RP of the polygonal mirror PM that performs the scanning of the point light SP (deflection of the light beam LBn) is set to every other one. In other words, the scanning of the point light SP by each scanning unit Un (U1 to U6) is repeated every other side (skipping one side) of the reflecting surface RP of the polygonal mirror PM. Therefore, during the period from when the scanning unit U1 scans the point light SP to when it performs the next scan, the light beams LB2 to LB6 can be sequentially allocated to each of the five scanning units U2 to U6 other than the scanning unit U1 to perform the scanning of the point light SP. That is, while the polygon mirror PM of a particular scanning unit Un (U1-U6) rotates two sides, light beams LB1-LB6 are distributed to each of the six scanning units Un (U1-U6), enabling all six scanning units Un (U1-U6) to scan the point light SP. In this case, the polygon mirror PM rotates two sides (90 degrees) between the time each scanning unit Un (U1-U6) starts scanning the point light SP and the time the next scanning unit Un (U1-U6) starts scanning the point light SP. To perform this drawing operation, the polygon mirrors PM of each of the six scanning units Un (U1-U6) are synchronously controlled to have the same rotation speed, and the angular positions of the reflection surfaces RP of each polygon mirror PM are synchronously controlled to have a predetermined phase relationship with each other.

Furthermore, since the reflection surface RP of the polygon mirror PM that performs the scanning of the point light SP (deflection of the light beam LBn) is set to every other surface, the number of scans of the point light SP along the drawing line SLn (SL1-SL6) becomes four during one rotation of the polygon mirror PM of each scanning unit Un (U1-U6). Therefore, compared to the case where the scanning of the point light SP (deflection of the light beam LBn) is repeated at each successive reflection surface RP of the polygon mirror PM, that is, compared to the case where the scanning of the point light SP (deflection of the light beam LBn) is performed at each reflection surface RP of the polygon mirror PM, the number of drawing lines SLn is halved. Therefore, it is preferable to also reduce the transport speed of the substrate FS by half. If it is not desired to reduce the transport speed of the substrate FS by half, the rotation speed and oscillation frequency Fs of the polygon mirror PM of each scanning unit Un (U1-U6) can be doubled. For example, when the point light SP is repeatedly scanned (the light beam LBn is deflected) on each continuous reflection surface RP of the polygonal mirror PM, the rotation speed of the polygonal mirror PM is 20,000 rpm and the oscillation frequency Fs of the light beam LB from the light source device 14' is 200 MHz. When the point light SP is repeatedly scanned (the light beam LBn is deflected) on every other reflection surface RP of the polygonal mirror PM, the rotation speed of the polygonal mirror PM is set to 40,000 rpm and the oscillation frequency Fs of the light beam LB from the light source device 14' is set to 400 MHz.

Here, the control device 18 manages which of the multiple scanning units Un (U1-U6) performs the scanning of the point light SP based on the origin signal SZ. However, since the origin sensor OPn of each scanning unit Un (U1-U6) generates the origin signal SZ when each reflection surface RP reaches a specified angular position, if this origin signal SZ is used directly, the control device 18 will determine that each scanning unit Un (U1-U6) scans the point light SP on each consecutive reflection surface RP. Therefore, between the time when one scanning unit Un performs the scanning of the point light SP and the time when the next scanning is performed, the light beam LBn cannot be distributed to the other five scanning units Un. Therefore, in order to set the reflection surface RP of the polygon mirror PM that performs the scanning of the point light SP to every other one, it is necessary to generate a sub-origin signal (sub-origin pulse signal) ZP that divides the origin signal SZ. Furthermore, as described above, since the origin signal SZ is detected using the preceding reflection surface RP in the rotational direction of the reflection surface RP that scans (deflects) the spot light SP, it is necessary to generate a sub-origin signal ZP that delays the generation timing of the origin signal SZ by a time Tpx. The following describes the structure of the sub-origin generation circuit CA that generates this sub-origin signal ZP.

Figure 31 is a block diagram of a sub-origin generation circuit CA for generating a sub-origin signal ZP with a generation timing delay of time Tpx by thinning out the origin signal SZ. Figure 32 is a timing chart showing the sub-origin signal ZP generated by the sub-origin generation circuit CA of Figure 31. The sub-origin generation circuit CA includes a frequency divider 330 and a delay circuit 332. The frequency divider 330 divides the frequency of the generation timing of the origin signal SZ into 1/2 and outputs the frequency as the origin signal SZ' to the delay circuit 332. The delay circuit 332 delays the received origin signal SZ' by time Tpx and outputs the result as the sub-origin signal ZP. Multiple sub-origin generation circuits CA are provided, corresponding to the origin sensors OPn of each scanning unit Un (U1 to U6).

In addition, the sub-origin generation circuit CA corresponding to the origin sensor OPn of the scanner unit Un may be represented by CAn. That is, the sub-origin generation circuit CA corresponding to the origin sensor OP1 of the scanner unit U1 may be represented by CA1, and the sub-origin generation circuits CA corresponding to the origin sensors OP2 to OP6 of the scanner units U2 to U6 may be represented by CA2 to CA6. Furthermore, the origin signal SZ output from the origin sensor OPn of the scanner unit Un may be represented by SZn. That is, the origin signal SZ output from the origin sensor OP1 of the scanner unit U1 may be represented by SZ1, and the origin signal SZ output from the origin sensors OP2 to OP6 of the scanner units U2 to U6 may be represented by SZ2 to SZ6. Furthermore, the origin signal SZ' and the sub-origin signal ZP generated based on the origin signal SZn may be represented by SZn' and ZPn, respectively. That is, there is a case where the origin signal SZ’ and sub-origin signal ZP generated based on the origin signal SZ1 are represented by SZ1’ and ZP1, and similarly, the origin signal SZ’ and sub-origin signal ZP generated based on the origin signals SZ2 to SZ6 are represented by SZ2’ to SZ6’ and ZP2 to ZP6.

FIG33 is a block diagram showing the electrical configuration of the exposure apparatus EX, and FIG34 is a timing chart showing the timing for outputting origin signals SZ1 to SZ6, sub-origin signals ZP1 to ZP6, and serial data DL1 to DL6. The control device 18 of the exposure apparatus EX includes a rotation control unit 350, a beam switching control unit 352, a drawing data output control unit 354, and an exposure control unit 356. Furthermore, the exposure apparatus EX includes motor drive circuits Drm1 to Drm6 that drive the polygon mirror drive unit RM, including the motors of the scanning units Un (U1 to U6).

The rotation control unit 350 controls the rotation of the polygon mirror PM of each scanning unit Un (U1 to U6) by controlling the motor drive circuits Drm1 to Drm6. The rotation control unit 350 controls the motor drive circuits Drm1 to Drm6 so that the rotation angle positions of the polygon mirrors PM of the multiple scanning units Un (U1 to U6) are in a predetermined phase relationship with each other, thereby causing the polygon mirrors PM of the multiple scanning units Un (U1 to U6) to rotate synchronously. In detail, the rotation control unit 350 controls the rotation of the polygon mirrors PM of the multiple scanning units Un (U1 to U6) so that the rotation speeds (number of rotations) of the polygon mirrors PM of the multiple scanning units U1 to U6 are the same and the phases of the rotation angle positions are shifted by a certain angle each time. In addition, the reference numerals PD1 to PD6 in Figure 33 represent control signals output from the rotation control unit 350 to the motor drive circuits Drm1 to Drm6.

In this fourth embodiment, the rotation speed Vp of the polygon mirror PM is set to 39,000 rpm (650 rps). In addition, since the number of reflection surfaces Np is set to 8, the scanning efficiency (α/β) is set to 1/3, and the reflection surface RP for scanning the point light SP is set to every other surface, the phase difference of the rotation angle position between the six polygon mirrors PM can be set to the maximum scanning rotation angle range α, that is, 15 degrees. The scanning of the point light SP is performed in the order of U1→U2→···→U6. Therefore, the rotation control unit 350 performs synchronization control in a manner such that the phase of the rotation angle position of each polygon mirror PM of the six scanning units U1 to U6 is rotated at a constant speed in this order while being shifted by 15 degrees each time. As a result, the phase deviation of the rotation angle position of the scanning unit U1 and the scanning unit U4 is exactly 45 degrees, which corresponds to the rotation angle of one surface. Therefore, the phase of the rotation angle position of the scanning unit U1 and the scanning unit U4, that is, the generation timing of the origin signals SZ1 and SZ4 can also be consistent. Similarly, the phase deviations of the rotation angle positions of scanning unit U2 and scanning unit U5, and the rotation angle positions of scanning unit U3 and scanning unit U6 are both 45 degrees. Therefore, the generation timings of the origin signals SZ2 and SZ5 from scanning unit U2 and scanning unit U5, respectively, and the generation timings of the origin signals SZ3 and SZ6 from scanning unit U3 and scanning unit U6, respectively, can also be consistent on the time axis.

Specifically, the rotation control unit 350 controls the rotation of the polygon mirrors PM of each scanning unit U1 to U6 via the motor drive circuits Drm1 to Drm6 so that the rotation of the polygon mirrors PM of the scanning units U1 and U4, the rotation of the polygon mirrors PM of the scanning units U2 and U5, and the rotation of the polygon mirrors PM of the scanning units U3 and U6 are each in a first control state. This first control state is a state in which the phase difference of the rotation pulse signal outputted each time the polygon mirror PM rotates once is 0 (zero). In other words, the rotation of the polygon mirrors PM of the scanning units U1 and U4 is controlled so that the phase difference of the rotation pulse signal outputted each time the polygon mirror PM rotates once is 0 (zero). Similarly, the rotation of the polygonal mirror PM of the scanning unit U2 and the scanning unit U5, and the scanning unit U3 and the scanning unit U6 is controlled in such a way that the phase difference of the rotation pulse signal outputted becomes 0 (zero) each time the polygonal mirror PM of the scanning unit U2 and the scanning unit U5, and the scanning unit U3 and the scanning unit U6 rotates once.

This rotation pulse signal can also be a signal that is output once every eight times the origin signal SZn of the scanning unit Un is output, via a frequency divider (not shown). Alternatively, the rotation pulse signal can be a signal output from an encoder (not shown) provided on the polygonal mirror drive unit RM of each scanning unit Un (U1-U6). A sensor for detecting the rotation pulse signal can also be provided near the polygonal mirror PM. In the example shown in FIG34 , a rotation pulse signal is generated every eight times the origin signal SZn of the scanning unit Un is output. The portion of the origin signal SZn corresponding to the generation of this rotation pulse signal is indicated by a dotted line. Furthermore, each origin signal SZ1 and each origin signal SZ4 are completely aligned in phase on the time axis, ignoring any error in the angle ηj formed between adjacent reflection surfaces RP (e.g., reflection surface RPa and reflection surface RPb) (see FIG29 ). Similarly, if the error in the angle ηj between adjacent reflection surfaces RP is not considered, the phases of the origin signals SZ2 and SZ5, and the origin signals SZ3 and SZ6 are all aligned on the time axis (see FIG29). Furthermore, FIG34 assumes that there is no error in the angle ηj between adjacent reflection surfaces RP for easier understanding.

Then, the rotation control unit 350 maintains the first control state and controls the rotation of the polygon mirrors PM of the scanning units U2 and U5 so that the phases of the rotation angle positions of the polygon mirrors PM of the scanning units U2 and U5 are offset by 15 degrees relative to the rotation angle positions of the polygon mirrors PM of the scanning units U1 and U4. Similarly, the rotation control unit 350 maintains the first control state and controls the rotation of the scanning units U3 and U6 so that the phases of the rotation angle positions of the polygon mirrors PM of the scanning units U3 and U6 are offset by 30 degrees relative to the rotation angle positions of the polygon mirrors PM of the scanning units U1 and U4. The time required for the polygon mirror PM to rotate by 15 degrees (the maximum scanning time of the light beam LBn) is set to Ts.

Specifically, the rotation control unit 350 controls the rotation of the polygon mirror PM of scanning units U2 and U5 so that the rotation pulse signals received by scanning units U2 and U5 are delayed by a time Ts relative to the rotation pulse signals received by scanning units U1 and U4 (see FIG34 ). Similarly, the rotation control unit 350 controls the rotation of the polygon mirror PM of scanning units U3 and U6 so that the rotation pulse signals received by scanning units U3 and U6 are delayed by a time 2×Ts relative to the rotation pulse signals received by scanning units U1 and U4 (see FIG34 ). If the rotation speed Vp of the polygon mirror PM is set to 39,000 rpm (650 rps), the time Ts is Ts = [1/(Vp×Np)]×(α/β)=1/(650×8×3) seconds (approximately 64.1 μ seconds). By controlling the rotation of the polygon mirror PM of each scanning unit U1 to U6 in this manner, the scanning units U1 to U6 can perform scanning of the spot light SP in a time-sharing manner in the order of U1 → U2 → ... → U6.

The beam switching control unit 352 controls the selective optical elements AOMn (AOM1-AOM6) of the beam switching component 20 to distribute the light beam LB from the light source device 14' to the six scanning units Un (U1-U6) from the time one scanning unit Un starts scanning until the time the next scanning unit Un starts scanning. Therefore, the beam switching control unit 352 repeatedly scans (deflects) the light beam LBn by the polygon mirror PM of each scanning unit Un (U1-U6) along every other reflective surface RP of the polygon mirror PM. This causes one of the light beams LB1-LB6 generated from the light beam LB to be incident on each scanning unit Un (U1-U6) in a time-sharing manner through the selective optical elements AOM1-AOM6.

Specifically, the beam switching control unit 352 includes a sub-origin generation circuit CAn (CA1-CA6) as shown in FIG. 31 , which generates a sub-origin signal ZPn (ZP1-ZP6) based on an origin signal SZn (SZ1-SZ6). When the sub-origin generation circuit CAn (CA1-CA6) generates the sub-origin signal ZPn (ZP1-ZP6), the selection optical elements AOMn (AOM1-AOM6) corresponding to the scanning units Un (U1-U6) are turned on for a predetermined time (on time Ton) in response to the generation of the sub-origin signal ZPn (ZP1-ZP6). For example, when the sub-origin signal ZP1 is generated, the selection optical element AOM1 corresponding to the scanning unit U1 is turned on for a predetermined time (on time Ton) in response to the generation of the sub-origin signal ZP1. This sub-origin signal ZPn is generated based on the origin signal SZn output from the origin sensor OPn. The frequency of the origin signal SZn is divided into 1/2, that is, the origin signal SZn is divided in half and delayed by a time delay of Tpx. This predetermined time (on time Ton) corresponds to the period from the time the sub-origin signal ZPn is generated to the time the sub-origin signal ZPn is generated by the scanning unit Un performing the next scan. In other words, it corresponds to the time Ts required for the polygon mirror PM to rotate 15 degrees. If the on time Ton of the selection optical element AOMn is set longer than time Ts, two of the selection optical elements AOMn will be in the on state simultaneously, making it impossible to correctly guide the light beams LB1 to LB6 to the scanning unit Un where the spot light SP is to be drawn. Therefore, the on time Ton is set to Ton ≤ Ts.

At this time, ignoring any errors in the angle ηj between adjacent reflection surfaces RP (e.g., reflection surface RPa and reflection surface RPb), each origin signal SZ1 and each origin signal SZ4 are synchronized on the time axis, with the phase difference between sub-origin signal ZP1 and sub-origin signal ZP4 being approximately half a cycle (see FIG34 ). This approximately half-cycle phase difference between sub-origin signal ZP1 and sub-origin signal ZP4 is caused by the frequency divider 330 of the sub-origin generation circuit CAn (CA1-CA6). Specifically, the frequency divider 330 causes the timing of thinning out the origin signal SZ1 to be offset from the timing of thinning out the origin signal SZ4 by approximately half a cycle.

Similarly, the relationship between the sub-origin signals ZP2 and ZP5 is set by the frequency divider 330 so that the phases of the sub-origin signals ZP2 and ZP5 are shifted by approximately half a cycle (see FIG34 ). Similarly, the relationship between the sub-origin signals ZP3 and ZP6 is set by the frequency divider 330 so that the phases of the sub-origin signals ZP3 and ZP6 are shifted by approximately half a cycle (see FIG34 ).

Therefore, as shown in Figure 34 , the generation timing of the sub-origin signals ZP1 to ZP6 generated by the scanning units U1 to U6 is staggered by a time interval of Ts. In the fourth embodiment, the order of the scanning units Un that scan the spot light SP is U1 → U2 → ... → U6. Therefore, when the sub-origin signal ZP2 is generated after a time interval of Ts following the generation of the sub-origin signal ZP1, the sub-origin signal ZPn is also generated at intervals of time Ts in the order of ZP1 → ZP2 → ... → ZP6. Therefore, the beam switching control unit 352 controls the selection optical elements AOMn (AOM1 to AOM6) of the beam switching unit 20 in accordance with the generated sub-origin signals ZPn (ZP1 to ZP6), thereby allowing the corresponding light beams LB1 to LB6 to be incident on each of the scanning units Un in the order of U1 → U2 → ... → U6. That is, the scanning (deflection) of the light beam LBn performed by the polygonal mirror PM of each scanning unit Un (U1~U6) is repeated on every other reflection surface RP of the polygonal mirror PM, thereby switching the light beam LBn incident on each scanning unit Un (U1~U6) in a time-sharing manner.

The drawing data output control unit 354 outputs one column of serial data DLn corresponding to the pattern of one drawing line SLn scanned by the spot light SP using the scanning unit Un as drawing bit string data Sdw to the driver circuit 206a of the light source device 14'. Since the order of the scanning units Un that scan the spot light SP is U1 → U2 → ... → U6, the drawing data output control unit 354 outputs one column of serial data DLn as the drawing bit string data Sdw repeatedly in the order of DL1 → DL2 → ... → DL6.

The structure of the drawing data output control unit 354 will be described in detail using FIG35 . The drawing data output control unit 354 includes six generation circuits 360, 362, 364, 366, 368, and 370, corresponding to scanning units U1 to U6, respectively, and an OR circuit GT8. Generation circuits 360 and 370 have the same structure. Specifically, generation circuit 360 includes a memory unit BM1, a counter unit CN1, and a gate unit GT1; generation circuit 362 includes a memory unit BM2, a counter unit CN2, and a gate unit GT2; generation circuit 364 includes a memory unit BM3, a counter unit CN3, and a gate unit GT3; generation circuit 366 includes a memory unit BM4, a counter unit CN4, and a gate unit GT4; generation circuit 368 includes a memory unit BM5, a counter unit CN5, and a gate unit GT5; and generation circuit 370 includes a memory unit BM6, a counter unit CN6, and a gate unit GT6. The configurations of the generating circuits 360 to 370 may be the same as those of the generating circuits 301 , 303 , and 305 shown in FIG. 16 .

The memory units BM1-BM6 store pattern data (bitmaps) corresponding to the patterns to be drawn and exposed by each scanning unit Un (U1-U6). The counter units CN1-CN6 are counters that output, for each pixel, serial data DL1-DL6 corresponding to the next drawing line SLn to be drawn, from the pattern data stored in the memory units BM1-BM6, in synchronization with the clock signal CLK. As shown in FIG34 , the counter units CN1-CN6 output a single serial data set DL1-DL6 after the sub-origin generation circuits CA1-CA6 of the beam switching control unit 352 output the sub-origin signals ZP1-ZP6.

The pattern data stored in each memory unit BM1-BM6 is shifted in the column direction by an address counter (not shown), etc., causing the output serial data DL1-DL6 to be shifted. Specifically, the columns read by the address counter (not shown) are shifted in the order of column 1, column 2, column 3, and so on. For example, in the case of the memory unit BM1 corresponding to the scanning unit U1, this shift is performed at the timing of the generation of the sub-origin signal ZP2 corresponding to the scanning unit U2, which is then performing the next scan, after the output of the serial data DL1 is completed. Similarly, the pattern data serial data DL2 stored in the memory unit BM2 is shifted at the timing of the generation of the sub-origin signal ZP3 corresponding to the scanning unit U3, which is then performing the next scan, after the output of the serial data DL2 is completed. Similarly, the pattern data serial data DL3-DL6 stored in the memory units BM3-BM6 is shifted at the timing of the generation of the sub-origin signals ZP4-ZP6 and ZP1 corresponding to the scanning units U4-U6 and U1, which are then performing the next scan, after the output of the serial data DL3-DL6 is completed. In addition, the scanning of the spot light SP is performed in the order of U1→U2→U3→···→U6.

In this manner, the serial data DL1-DL6 output sequentially are applied to a six-input OR circuit GT8 via gates GT1-GT6, which are opened for a predetermined time (On time Ton) after the sub-origin signals ZP1-ZP6 are applied. OR circuit GT8 repeatedly synthesizes serial data DLn in the order of serial data DL1 → DL2 → DL3 → DL4 → DL5 → DL6 → DL1, etc., and outputs this serial data DLn as drawing bit string data Sdw to the driver circuit 206a of the light source device 14'. In this manner, each scanning unit Un (U1-U6) can simultaneously draw and expose a pattern corresponding to the pattern data while scanning with the spot light SP.

In this fourth embodiment, pattern data is prepared for each scanning unit Un (U1 to U6), and serial data DL1 to DL6 are output from the pattern data of each scanning unit Un (U1 to U6) in the order in which the scanning units Un scan the spot light SP. However, since the order in which the scanning units Un scan the spot light SP is predetermined, a single pattern data set can be prepared by combining the serial data DL1 to DL6 of the pattern data of each scanning unit Un (U1 to U6). In other words, a single pattern data set can be constructed in which the serial data DLn (DL1 to DL6) of each column of pattern data of each scanning unit Un (U1 to U6) is arranged in accordance with the order in which the scanning units Un scan the spot light SP. In this case, it is sufficient to output the serial data DLn of the pattern data in order, starting from the first column, based on the sub-origin signal ZPn (ZP1 to ZP6) from the origin sensor OPn of each scanning unit Un (U1 to U6).

In addition, the exposure control unit 356 shown in Figure 33 is used to control the rotation control unit 350, the beam switching control unit 352, the drawing data output control unit 354, etc. The exposure control unit 356 analyzes the imaging signal ig (ig1 to ig4) obtained by the alignment microscope AMm (AM1 to AM4) to detect the position of the alignment mark MKm (MK1 to MK4) on the substrate FS. Then, based on the position of the detected alignment mark MKm (MK1 to MK4), the exposure control unit 356 detects (determines) the starting position of the drawing exposure of the exposure area W on the substrate FS. The exposure control unit 356 has a counter circuit 356a, which counts the detection signals detected by the encoders EN1a to EN3a and EN1b to EN3b shown in Figure 24. The exposure control unit 356 determines whether the start position of the drawing exposure on the substrate FS is located on the drawing lines SL1, SL3, or SL5 based on the count values (mark detection position) of encoders EN1a and EN1b when the start position of the drawing exposure is detected, and the count values (position of odd-numbered drawing lines SLn) of encoders EN2a and EN2b. If the exposure control unit 356 determines that the start position of the drawing exposure is located on the drawing lines SL1, SL3, or SL5, it controls the drawing data output control unit 354 to cause the scanning units U1, U3, and U5 to begin scanning with the spot light SP. Furthermore, under the control of the exposure control unit 356, the rotation control unit 350 and the beam switching control unit 352 control the rotation of the polygon mirror PM of each scanning unit Un (U1-U6) and the distribution of the light beam LBn by the beam switching unit 20 based on the rotation pulse signal and the sub-origin signal ZPn (ZP1-ZP6).

The exposure control unit 356 determines whether the start position of the drawing exposure of the substrate FS is located on the drawing lines SL2, SL4, and SL6 based on the count values (mark detection positions) of the encoders EN1a and EN1b when the start position of the drawing exposure is detected, and the count values (positions of even-numbered drawing lines) of the encoders EN3a and EN3b. If the exposure control unit 356 determines that the start position of the drawing exposure is located on the drawing lines SL2, SL4, and SL6, it controls the drawing data output control unit 354 to cause the scanning units U2, U4, and U6 to start scanning with the spot light SP.

As shown in FIG. 25 , the drawing lines SL1, SL3, and SL5 are first exposed according to the conveying direction (+X direction) of the substrate FS. After the substrate FS is conveyed a predetermined distance, the drawing lines SL2, SL4, and SL6 are then exposed. On the other hand, since the polygonal mirrors PM of the six scanning units U1 to U6 are rotationally controlled to maintain a certain angular phase with each other, the sub-origin signals ZP1 to ZP6 are continuously generated in sequence with a phase difference of time Ts as shown in FIG. 34 . Therefore, from the start time of the drawing exposure of the drawing lines SL1, SL3, and SL5 to the time immediately before the start of the drawing exposure of the drawing lines SL2, SL4, and SL6, the gates GT2, GT4, and GT6 in FIG. 35 are also opened by the sub-origin signals ZP2, ZP4, and ZP6, and the action of selecting the optical elements AOM2, AOM4, and AOM6 to be in the On state for a certain time Ton is repeatedly performed. Therefore, in the configuration of FIG. 33 , a selection gate circuit may be provided within the beam switching control unit 352. This selection gate circuit selects whether to transmit or prohibit transmission of each of the generated sub-origin signals ZP1 to ZP6 to the drawing data output control unit 354 based on the count values of encoders EN1a and EN1b or the count values of encoders EN2a and EN2b determined by the exposure control unit 356. Furthermore, the sub-origin signals ZP1 to ZP6 may be supplied to the driver circuits DRVn (DRV1 to DRV6) (see FIG. 38 ) of the selection optical elements AOM1 to AOM6 corresponding to the scanning units U1 to U6, respectively, via this selection gate circuit.

As described above, since the drawing lines SL1, SL3, and SL5 are located upstream of the drawing lines SL2, SL4, and SL6 in the conveyance direction of the substrate FS, the start position of the drawing exposure of the exposure area W of the substrate FS first reaches the drawing lines SL1, SL3, and SL5, and then reaches the drawing lines SL2, SL4, and SL6 after a certain period of time. Therefore, before the start position of the drawing exposure reaches the drawing lines SL2, SL4, and SL6, the pattern drawing exposure is performed only by the scanning units U1, U3, and U5. Therefore, without the selection gate circuits for the sub-origin signals ZP1 to ZP6 described above being provided in the beam switching control unit 352, the exposure control unit 356 sets all pixel data corresponding to the serial data DL2, DL4, and DL6 in the drawing bit string data Sdw output to the driver circuit 206a of the light source device 14' to low "(0)", thereby effectively canceling the drawing exposure performed by the scanning units U2, U4, and U6. During the cancel period, the columns of serial data DL2, DL4, and DL6 output from the memory units BM2, BM4, and BM6 remain unchanged and remain at column 1. Then, after the start position of the drawing exposure in the exposure area W reaches the drawing lines SL2, SL4, and SL6, the output of the serial data DL2, DL4, and DL6 begins, and the serial data DL2, DL4, and DL6 are shifted in the column direction.

Similarly, the end position of the drawing exposure in the exposure area W first reaches the drawing lines SL1, SL3, and SL5, and then reaches the drawing lines SL2, SL4, and SL6 after a certain period of time. Therefore, after the end position of the drawing exposure reaches the drawing lines SL1, SL3, and SL5 and before reaching the drawing lines SL2, SL4, and SL6, the drawing exposure of the pattern is performed only by the scanning units U2, U4, and U6. Therefore, when the selection gate circuits for the sub-origin signals ZP1 to ZP6 described above are not provided in the beam switching control unit 352, the exposure control unit 356 sets all the pixel data corresponding to the serial data DL1, DL3, and DL5 in the drawing bit string data Sdw output to the driver circuit 206a of the light source device 14' to low "(0)", thereby substantially canceling the drawing exposure by the scanning units U1, U3, and U5. In addition, when a selection gate circuit is not provided, even when the drawing exposure is canceled, the optical elements AOM1, AOM3, AOM5 are selected to repeatedly perform the action of selectively becoming the On state for a certain period of time Ton in response to the sub-origin signals ZP1, ZP3, ZP5 in such a manner that the light beams LB1, LB3, LB5 are introduced into the scanning units U1, U3, U5 in which the drawing exposure is canceled.

As described above, in this fourth embodiment, the beam switching control unit 352 controls the beam switching component 20 to repeatedly deflect (scan) the polygonal mirror PM at every other reflective surface RP of the scanning unit Un (U1-U6), causing each of the multiple scanning units Un (U1-U6) to sequentially perform one-dimensional scanning of the point light SP. This allows a single light beam LB to be distributed to multiple scanning units Un (U1-U6) without shortening the length of the traced line SLn (SL1-SL6) scanned by the point light SP, effectively utilizing the light beam LB. Furthermore, since the polygonal shape (polygonal shape) of the polygonal mirror PM can be made approximately circular, the rotation speed of the polygonal mirror PM can be prevented from decreasing, allowing the polygonal mirror PM to rotate at high speed.

The beam switching unit 20 includes n selection optical elements AOMn (AOM1 to AOM6) arranged in series along the direction of travel of the light beam LB from the light source device 14'. These select one of the n light beams LBn diffracted and deflected, and direct it to the corresponding scanning unit Un. This allows for easy selection of the scanning unit Un (U1 to U6) on which the light beam LBn is to be incident, enabling efficient concentration of the light beam LB from the light source device 14' on the scanning unit Un to perform drawing exposure, resulting in a high exposure value. For example, when the light beam LB emitted from the light source device 14' is amplitude-split into six light beams using multiple beam splitters, and each of the six split light beams LBn (LB1 to LB6) is introduced into the six scanning units U1 to U6 via a drawing acousto-optic modulator element modulated according to the drawing data serial data DL1 to DL6, if the attenuation of the beam intensity in the drawing acousto-optic modulator element is set to 20% and the attenuation of the beam intensity in the scanning unit Un is set to 30%, the intensity of the spot light SP in one scanning unit Un is approximately 9.3% of the original intensity of the light beam LB, when the intensity of the original light beam LB is 100%. On the other hand, when the light beam LB from the light source device 14' is deflected by the selection optical element AOMn and incident on one of the six scanning units Un, when the attenuation of the beam intensity in the selection optical element AOMn is set to 20%, the intensity of the spot light SP in one scanning unit Un is approximately 56% of the original intensity of the light beam LB.

The rotation control unit 350 controls the rotation of the polygon mirrors PM of the plurality of scanning units Un (U1-U6) so that the rotation speeds are the same and the phases of the rotational angle positions are shifted by a certain angle each time. Thus, between one-dimensional scanning of the spot light SP by one scanning unit Un and the next one-dimensional scanning, one-dimensional scanning of the spot light SP by the other plurality of scanning units Un can be performed sequentially.

In addition, in the fourth embodiment described above, a method of distributing one light beam LB to six scanning units Un was described, but one light beam LB from the light source device 14' may also be distributed to nine scanning units Un (U1 to U9). In this case, if the scanning efficiency (α/β) of the polygonal mirror PM is set to 1/3, the light beam LBn can be distributed to the nine scanning units U1 to U9 while the polygonal mirror PM rotates by the amount of three reflecting surfaces RP, so that the point light SP is scanned on every two reflecting surfaces RP. Thus, from the time when the point light SP is scanned by one scanning unit Un to the time when the next point light SP is scanned, the other eight scanning units Un can scan the point light SP in sequence. In addition, if the scanning efficiency of the polygonal mirror PM is set to 1/3, the polygonal mirror PM can rotate by the amount of three reflecting surfaces RP to distribute one light beam LB to the nine scanning units Un. Therefore, the frequency divider 330 of the sub-origin generating circuit CAn divides the frequency of the generation timing of the origin signal SZn into 1/3. In this case, the convolution pulse signals of scanning units U1, U4, and U7 are synchronized (in phase on the time axis). Similarly, the convolution pulse signals of scanning units U2, U5, and U8 are synchronized, and the convolution pulse signals of scanning units U3, U6, and U9 are synchronized. Furthermore, the convolution pulse signals of scanning units U2, U5, and U8 are generated with a delay of time Ts relative to the convolution pulse signals of scanning units U1, U4, and U7, while the convolution pulse signals of scanning units U3, U6, and U9 are generated with a delay of 2×Ts relative to the convolution pulse signals of scanning units U1, U4, and U7. Furthermore, the phases of the secondary origin signals ZP1, ZP4, and ZP7 generated by the scanning units U1, U4, and U7 are shifted by one-third of a cycle. Similarly, the phases of the secondary origin signals ZP2, ZP5, and ZP8 generated by the scanning units U2, U5, and U8, and the phases of the secondary origin signals ZP3, ZP6, and ZP9 generated by the scanning units U3, U6, and U9 are also shifted by one-third of a cycle. Furthermore, the time Ts is the time for the polygon mirror PM to rotate within the scanning rotation angle range α' that enables scanning of the spot light SP. The value obtained by multiplying the angle β, which is the rotation of the polygon mirror PM by one reflecting surface RP, by the scanning efficiency is the scanning rotation angle range α'.

When the scanning efficiency of the polygon mirror PM is set to 1/3 and one light beam LB is distributed to the 12 scanning units Un (U1 to U12), the light beam LBn can be distributed to the 12 scanning units U1 to U12 while the polygon mirror PM rotates by the amount of four reflection surfaces RP. Therefore, the scanning of the point light SP is performed on every third reflection surface RP. In addition, if the scanning efficiency of the polygon mirror PM is set to 1/3, the polygon mirror PM can rotate by the amount of four reflection surfaces RP to cause the light beam LBn (LB1 to LB12, which is a light beam obtained by selectively deflecting the light beam LB from the light source device 14' through 12 selection optical elements AOMn (AOM1 to AOM12) arranged in series) to be incident on a corresponding scanning unit Un (U1 to U12). Therefore, the frequency divider 330 of the sub-origin generating circuit CAn divides the frequency of the generation timing of the origin signal SZn into 1/4. In this case, the convolution pulse signals of scanning units U1, U4, U7, and U10 are synchronized (in phase on the time axis). Similarly, the convolution pulse signals of scanning units U2, U5, U8, and U11 are synchronized, and the convolution pulse signals of scanning units U3, U6, U9, and U12 are synchronized. Furthermore, the convolution pulse signals of scanning units U2, U5, U8, and U11 are generated with a delay of time Ts relative to the convolution pulse signals of scanning units U1, U4, U7, and U10, while the convolution pulse signals of scanning units U3, U6, U9, and U12 are generated with a delay of 2×Ts relative to the convolution pulse signals of scanning units U1, U4, U7, and U10. In addition, the phases of the generation timings of the sub-origin signals ZP1, ZP4, ZP7, and ZP10 of the scanning units U1, U4, U7, and U10 are shifted by 1/4 cycle one by one. Similarly, the phases of the generation timings of the sub-origin signals ZP2, ZP5, ZP7, and ZP11 of the scanning units U2, U5, U8, and U11 and the phases of the generation timings of the sub-origin signals ZP3, ZP6, ZP9, and ZP12 of the scanning units U3, U6, U9, and U12 are also shifted by 1/4 cycle one by one.

In addition, in the fourth embodiment described above, the scanning efficiency of the polygonal mirror PM of the scanning unit Un is described as 1/3, but the scanning efficiency may be 1/2 or 1/4. When the scanning efficiency is 1/2, the light beam LBn can be distributed to two scanning units Un while the polygonal mirror PM rotates by the amount of one reflecting surface RP. Therefore, when one light beam LBn is to be distributed to six scanning units Un, the scanning of the point light SP is performed on every two reflecting surfaces RP of the polygonal mirror PM. That is, when the scanning efficiency of the polygonal mirror PM is 1/2, the light beam LBn can be distributed to six scanning units Un while the polygonal mirror PM rotates by the amount of three reflecting surfaces RP. Thus, from the time when the point light SP is scanned by one scanning unit Un to the time when the next scanning of the point light SP is performed, the other five scanning units Un can scan the point light SP in sequence. Furthermore, if the scanning efficiency of the polygon mirror PM is set to 1/2, the polygon mirror PM can rotate by the amount of three reflecting surfaces RP to distribute one light beam LB to the six scanning units Un. Therefore, the frequency divider 330 of the sub-origin generating circuit CAn divides the frequency of the generation timing of the origin signal SZn into 1/3. In this case, the rotation pulse signals of the scanning units U1, U3, and U5 are synchronized. Similarly, the rotation pulse signals of the scanning units U2, U4, and U6 are synchronized. Moreover, the rotation pulse signals of the scanning units U2, U4, and U6 are generated with a delay time Ts relative to the rotation pulse signals of the scanning units U1, U3, and U5. In addition, the phases of the generation timings of the sub-origin signals ZP1, ZP3, and ZP5 of the scanning units U1, U3, and U5 are shifted by 1/3 of a cycle, and the phases of the generation timings of the sub-origin signals ZP2, ZP4, and ZP6 of the scanning units U2, U4, and U6 are also shifted by 1/3 of a cycle.

When the scanning efficiency of the polygon mirror PM is 1/4, the light beam LBn can be distributed to four scanning units Un while the polygon mirror PM rotates by the amount of one reflecting surface RP. Therefore, when one light beam LB is to be distributed to eight scanning units Un, the scanning of the point light SP is performed on every other reflecting surface RP of the polygon mirror PM. That is, when the scanning efficiency of the polygon mirror PM is 1/4, the light beam LBn can be distributed to eight scanning units Un while the polygon mirror PM rotates by the amount of two reflecting surfaces RP. Thus, from the time when the point light SP is scanned by one scanning unit Un to the time when the next point light SP is scanned, the other seven scanning units Un can scan the point light SP in sequence. In addition, if the scanning efficiency of the polygon mirror PM is set to 1/4, the light beam LB can be distributed to eight scanning units Un while the polygon mirror PM rotates by the amount of two reflecting surfaces RP. Therefore, the frequency divider 330 of the sub-origin generating circuit CAn divides the frequency of the generation timing of the origin signal SZn into 1/2. In this case, the rotation pulse signals of scanning units U1 and U5 are synchronized, as are the rotation pulse signals of scanning units U2 and U6. Similarly, the rotation pulse signals of scanning units U3 and U7 are synchronized, and the rotation pulse signals of scanning units U4 and U8 are synchronized. Furthermore, the rotation pulse signals of scanning units U2 and U6 are generated with a delay of only time Ts relative to those of scanning units U1 and U5. The rotation pulse signals of scanning units U3 and U7 are generated with a delay of 2 times time Ts relative to those of scanning units U1 and U5, and the rotation pulse signals of scanning units U4 and U8 are generated with a delay of 3 times time Ts relative to those of scanning units U1 and U5. Furthermore, the phases of the generation timings of the auxiliary origin signals ZP1 and ZP5 of scanning units U1 and U5 are shifted by 1/2 cycle, and the phases of the generation timings of the auxiliary origin signals ZP2 and ZP6 of scanning units U2 and U6 are also shifted by 1/2 cycle. Similarly, the phases of the generation timings of the sub-origin signals ZP3 and ZP7 of the scanning units U3 and U7 and the phases of the generation timings of the sub-origin signals ZP4 and ZP8 of the scanning units U4 and U8 are also shifted by 1/2 cycle.

Furthermore, in the fourth embodiment, the polygonal mirror PM is octagonal (having eight reflecting surfaces RP). However, it may also be a hexagon, a heptagon, or even a pentagon or larger. This will affect the scanning efficiency of the polygonal mirror PM. Generally, the greater the number of reflecting surfaces Np of the polygonal mirror PM, the greater the scanning efficiency of each reflecting surface RP of the polygonal mirror PM. The smaller the number of reflecting surfaces Np, the lower the scanning efficiency of the polygonal mirror PM.

The maximum scanning rotation angle range α of the polygonal mirror PM that can project the point light SP onto the substrate FS for scanning is determined by the incident angle of the fθ lens FT (equivalent to the scanning angle range θs in FIG29 ). Therefore, the polygonal mirror PM with the optimal number of reflection surfaces Np can be selected corresponding to the incident angle. In the case of the fθ lens FT with an incident angle (θs) less than 30 degrees as in the previous example, it is also possible to use a 24-faceted polygonal mirror PM in which the reflection surface RP is changed by rotating half of it, that is, 15 degrees, or a 12-faceted polygonal mirror PM in which the reflection surface RP is changed by rotating 30 degrees. In this case, the scanning efficiency (α/β) in the 24-faceted polygonal mirror PM is greater than 1/2 and less than 1.0, so the 24-faceted polygonal mirror PM of each of the six scanning units U1 to U6 is controlled to skip five surfaces to scan the point light SP. In addition, since the scanning efficiency of the 12-face polygon mirror PM is greater than 1/3 and less than 1/2, the 12-face polygon mirror PM of each of the six scanning units U1 to U6 is controlled so as to scan the spot light SP while skipping two surfaces.

[Fifth embodiment]

In the fourth embodiment described above, the scanning (deflection) of the point light SP is always repeated for each reflection surface RP of the polygon mirror PM. However, in the fifth embodiment, the scanning (deflection) of the point light SP can be arbitrarily switched to a first state in which the scanning is repeated for each successive reflection surface RP of the polygon mirror PM, or a second state in which the scanning is repeated for each reflection surface RP of the polygon mirror PM. In other words, between the time when the scanning unit U1 starts scanning the point light SP and the time before the next scan begins, the light beam LB can be switched to be distributed to the three scanning units Un in a time-sharing manner, or to the six scanning units Un in a time-sharing manner.

Since the scanning efficiency of the polygon mirror PM is 1/3, when the scanning of the point light SP is repeated on each continuous reflection surface RP of the polygon mirror PM, for example, during the period from when the scanning unit U1 scans the point light SP to when the next scan is performed, the light beam LB can only be distributed to two scanning units Un other than the scanning unit U1. Therefore, two light beams LB are prepared, and the first light beam LB is distributed to the three scanning units Un in a time-sharing manner, and the second light beam LB is distributed to the remaining three scanning units Un in a time-sharing manner. Therefore, the scanning of the point light SP is performed in parallel by the two scanning units Un. The two light beams LB can be generated by providing two light source devices 14', or by dividing the light beam LB from one light source device 14' by a beam splitter or the like. In the exposure device EX of the fifth embodiment shown in Figures 36 to 40, two light source devices 14' (14A', 14B') are provided (refer to Figure 38). In addition, in the fifth embodiment, the same figure marks are given to the same structures as the fourth embodiment, and only the different parts are described.

FIG36 is a structural diagram of a beam switching component (beam delivery unit) 20A according to the fifth embodiment. Similar to the beam switching component 20 in FIG26 , the beam switching component 20A includes multiple selection optical elements AOMn (AOM1 to AOM6), multiple focusing lenses CD1 to CD6, multiple reflectors M1 to M12, multiple mirrors IM1 to IM6, and multiple collimating lenses CL1 to CL6. Furthermore, it includes reflectors M13 and M14 and absorbers TR1 and TR2. The absorber TR1 corresponds to the absorber TR in FIG26 shown in the fourth embodiment, and absorbs the light beam LB reflected by the reflector M12.

The selective optical elements AOM1 to AOM3 constitute an optical element module (first optical element module) OM1, and the selective optical elements AOM4 to AOM6 constitute an optical element module (second optical element module) OM2. As described in the fourth embodiment above, the selective optical elements AOM1 to AOM3 of the first optical element module OM1 are arranged in a row along the direction of travel of the light beam LB. Similarly, the selective optical elements AOM4 to AOM6 of the second optical element module OM2 are also arranged in a row along the direction of travel of the light beam LB. In addition, the scanning units U1 to U3 corresponding to the selective optical elements AOM1 to AOM3 of the first optical element module OM1 are set as the first scanning module. In addition, the scanning units U4 to U6 corresponding to the selective optical elements AOM4 to AOM6 of the second optical element module OM2 are set as the second scanning module. The scanning units U1 to U3 of the first scanning module and the scanning units U4 to U6 of the second scanning module are arranged in a predetermined arrangement relationship as described in the fourth embodiment above.

In the fifth embodiment, the reflectors M6, M13, and M14 constitute a configuration switching member (movable member) SWE that can be switched between a first configuration state and a second configuration state. In the first configuration state, the first optical element module OM1 and the second optical element module OM2 are arranged in parallel in the direction of travel of the light beam LB. In the second configuration state, the first optical element module OM1 and the second optical element module OM2 are arranged in series in the direction of travel of the light beam LB. The configuration switching member SWE includes a sliding member SE that supports the reflectors M6, M13, and M14. The sliding member SE is movable in the X direction relative to the supporting member IUB. The movement of the sliding member SE (configuration switching member SWE) in the X direction is performed by an actuator AC (see FIG. 38 ). The actuator AC is driven by the drive control unit 352a (see FIG. 38 ) of the beam switching control unit 352.

In the first configuration state, the light beams LB from the two light source devices 14' (14A', 14B') are incident on the first optical element module OM1 and the second optical element module OM2 in parallel. In the second configuration state, the light beam LB from the single light source device 14' (14A') is incident on both the first optical element module OM1 and the second optical element module OM2. In other words, in the second configuration state, the light beam LB transmitted through the first optical element module OM1 is incident on the second optical element module OM2. Figure 36 shows the second configuration state, in which the first optical element module OM1 and the second optical element module OM2 are arranged in series by configuring the switching element SWE. In other words, in this second configuration state, all the selective optical elements AOM1 to AOM6 of the first optical element module OM1 and the second optical element module OM2 are arranged in series along the direction of travel of the light beam LB, similar to Figure 26 shown in the fourth embodiment described above. Therefore, similarly to the fourth embodiment described above, a scanning unit Un on which a deflected light beam LBn is incident can be selected from the first scanning module and the second scanning module (U1 to U6) by each selection optical element AOMn (AOM1 to AOM6) of the first optical element module OM1 and the second optical element module OM2 arranged in series. In addition, the position of the configuration switching component SWE in Figure 36 is referred to as the second position. In addition, in the first configuration state, the light beam LB incident on the first optical element module OM1 (AOM1 to AOM3) is referred to as the light beam LBa from the first light source device 14A', and in the first configuration state, the light beam incident on the second optical element module OM2 (AOM4 to AOM6) is referred to as the light beam LBb from the second light source device 14B'.

When the configuration switching member SWE moves toward the -X direction and reaches the first position, the first optical element module OM1 and the second optical element module OM2 enter a first configuration state in which they are arranged side by side. Figure 37 shows the optical paths of the light beams LBa and LBb when the configuration switching member SWE is in the first position. In the first configuration state, the light beam LBa is incident on the first optical element module OM1, and the light beam LBb is incident on the second optical element module OM2. To distinguish between the light beams LB incident on the first optical element module OM1 and the second optical element module OM2, the light beam LB incident on the first optical element module OM1 is represented by LBa, and the light beam LB directly incident on the second optical element module OM2 is represented by LBb.

As shown in Figure 37, when the configuration switching member SWE is moved to the first position, the position of the reflector M6 shifts in the -X direction. Therefore, the light beam LBa reflected by the reflector M6 does not enter the reflector M7 but enters the absorber TR2. Therefore, the light beam LBa from the first light source unit 14A' entering the first optical element module OM1 enters only the first optical element module OM1 (selective optical elements AOM1-AOM3) and does not enter the second optical element module OM2. In other words, the light beam LBa can only pass through the selective optical elements AOM1-AOM3. Furthermore, when the configuration switching member SWE is in the first position, the light beam LBb emitted from the second light source unit 14B' and traveling in the +Y direction toward the reflector M13 is guided by the reflectors M13 and M14 toward the reflector M7. Therefore, the light beam LBb can only pass through the second optical element module OM2 (selective optical elements AOM4-AOM6).

Therefore, the first optical element module OM1 can cause one of the light beams LB1 to LB3 deflected from the light beam LBa to enter one of the three scanning units U1 to U3 constituting the first scanning module via the three selection optical elements AOM1 to AOM3 arranged in series. Furthermore, the second optical element module OM2 can cause one of the light beams LB4 to LB6 deflected from the light beam LBb to enter one of the three scanning units U4 to U6 constituting the second scanning module via the three selection optical elements AOM4 to AOM6 arranged in series. Therefore, the first optical element module OM1 (AOM1 to AOM3) and the second optical element module OM2 (AOM4 to AOM6) arranged in parallel can each select a scanning unit Un from the first scanning module (U1 to U3) and the second scanning module (U4 to U6) on which the light beam LB is to enter. In this case, an exposure operation is performed in parallel by one of the scanning units Un in the first scanning module and one of the scanning units Un in the second scanning module, in which the spot light SP is scanned along the drawing line SLn.

In the first state (first drawing mode) in which the scanning (deflection) of the spot light SP is repeated for each successive reflection surface RP of the polygon mirror PM, the beam switching control unit 352 controls the actuator AC to place the configuration switching element SWE in the first position. Furthermore, in the second state (second drawing mode) in which the beam switching control unit 352 is repeated for each reflection surface RP of the polygon mirror PM, the actuator AC is controlled to place the configuration switching element SWE in the second position.

FIG38 shows the configuration of the light beam switching control unit 352 in the fifth embodiment. FIG38 also illustrates the selective optical elements AOM1 to AOM6 and the light source device 14' (14A', 14B') that are controlled by the light beam switching control unit 352. The light source device 14' that directs the light beam LBa from the first optical element module OM1 is indicated by 14A', while the light source device 14' that directs the light beam LBb only to the second optical element module OM2 is indicated by 14B'.

When the configuration switching member SWE is in the second position, as shown in FIG38 , the light beam LBa (LB) from the light source device 14A′ can pass through (be transmitted by) the selection optical element AOMn in the order of AOM1 → AOM2 → AOM3 → ... → AOM6, and the light beam LBa that has passed through the selection optical element AOM6 is incident on the absorber TR1. Furthermore, when the configuration switching member SWE is moved to the first position, the light beam LBa from the light source device 14A′ can pass through the selection optical element AOMn in the order of AOM1 → AOM2 → AOM3, and the light beam LBa that has passed through the selection optical element AOM3 is incident on the absorber TR2. Furthermore, when the configuration switching member SWE is moved to the first position, the light beam LBb from the light source device 14B′ can pass through the selection optical element AOMn in the order of AOM4 → AOM5 → AOM6, and the light beam LB that has passed through the selection optical element AOM6 is incident on the absorber TR1. The configuration switching member SWE in FIG38 is a conceptual diagram and differs from the actual structure of the configuration switching member SWE shown in FIG36 and FIG37 . In the example shown in FIG38 , the configuration switching member SWE is in its second position, i.e., in the second configuration state in which the first optical element module OM1 and the second optical element module OM2 are arranged in series, and the selective optical element AOM5 is in the "on" state. Consequently, the light beam LBa from the light source device 14A' is diffracted and deflected, resulting in a light beam LB5 incident on the scanning unit U5.

The beam switching control unit 352 includes driver circuits DRVn (DRV1-DRV6) that drive each of the selection optical elements AOM1-AOM6 with ultrasonic (high-frequency) signals, and sub-origin generation circuits CAan (CAa1-CAa6) that generate sub-origin signals ZPn (ZP1-ZP6) based on origin signals SZn (SZ1-SZ6) from origin sensors OPn of the respective scanning units Un (U1-U6). The exposure control unit 356 transmits information about the on-time Ton, which turns the selection optical elements AOM1-AOM6 on for a predetermined period of time after receiving the sub-origin signals ZPn (ZP1-ZP6). Upon receiving the sub-origin signal ZP1 from the sub-origin generation circuit CAa1, the driver circuit DRV1 turns the selection optical element AOM1 on for the on-time Ton. Similarly, driver circuits DRV2-DRV6, upon receiving sub-origin signals ZP2-ZP6 from sub-origin generation circuits CAa2-CAa6, turn on the selection optical elements AOM2-AOM6 for an on-time Ton. When the exposure control unit 356 changes the rotational speed of the polygon mirror PM, it adjusts the length of the on-time Ton accordingly. Driver circuits DRVn (DRV1-DRV6) are also provided in the beam switching control unit 352 shown in FIG. 33 in the fourth embodiment.

The sub-origin generating circuit CAan (CAa1 to CAa6) includes a logic circuit LCC and a delay circuit 332. The logic circuit LCC of the sub-origin generating circuit CAan (CAa1 to CAa6) receives an input of the origin signal SZn (SZ1 to SZ6) from the origin sensor OPn of each scanning unit Un (U1 to U6). That is, the logic circuit LCC of the sub-origin generating circuit CAa1 receives an input of the origin signal SZ1, and similarly, the logic circuit LCC of the sub-origin generating circuits CAa2 to CAa6 receives an input of the origin signal SZ2 to SZ6. Furthermore, the logic circuit LCC of each sub-origin generating circuit CAan (CAa1 to CAa6) receives an input of the state signal STS. This state signal (logical value) STS is set to "1" in the case of the first state, which is repeated for each successive reflection surface RP of the polygon mirror PM, and is set to "0" in the case of the second state, which is repeated for each reflection surface RP of the polygon mirror PM. This state signal STS is transmitted from the exposure control unit 356 .

Each logic circuit LCC generates an origin signal SZn' (SZ1' to SZ6') based on the input origin signal SZn (SZ1 to SZ6), and outputs it to each delay circuit 332. Each delay circuit 332 delays the input origin signal SZn' (SZ1' to SZ6') by a time Tpx and outputs a sub-origin signal ZPn (ZP1 to ZP6).

Figure 39 shows the structure of a logic circuit LCC that receives inputs of origin signal SZn (SZ1-SZ6) and state signal STS. Logic circuit LCC consists of a two-input OR gate LC1, a two-input AND gate LC2, and a single-shot pulse generator LC3. State signal STS is applied as one input to OR gate LC1. The output signal (logical value) of OR gate LC1 is applied as one input to AND gate LC2, and origin signal SZn is applied as the other input to AND gate LC2. The output signal (logical value) of AND gate LC2 is input to delay circuit 332 as origin signal SZn'. Single-shot pulse generator LC3 normally outputs signal SDo with a logical value of "1." However, when origin signal SZn' (SZ1'-SZ6') is generated, it outputs signal SDo with a logical value of "0" only for a certain period of time, Tdp. That is, when generating origin signals SZn' (SZ1' to SZ6'), the single-shot pulse generator LC3 inverts the logic value of the signal SDo for a certain time Tdp. The time Tdp is set to satisfy the relationship 2×Tpx>Tdp>Tpx, preferably Tdp≈1.5×Tpx.

FIG40 is a diagram illustrating the timing of the operation of the logic circuit LCC of FIG39 . The left half of FIG40 illustrates a first state in which the scanning units Un (U1-U6) scan the spot light SP without skipping over each successive reflection surface RP. The right half illustrates a second state in which the scanning units Un (U1-U6) scan the spot light SP without skipping over a single reflection surface RP. Furthermore, FIG40 assumes, for ease of understanding, that the angle ηj formed between adjacent reflection surfaces RP (e.g., reflection surface RPa and reflection surface RPb) of the polygon mirror PM is consistent, and that the origin signal SZn is accurately generated at intervals of time Tpx.

In the first state in which the scanning of the point light SP is performed on each reflecting surface RP without skipping, the state signal STS is "1", and therefore, the output signal of the OR gate LC1 is always "1" regardless of the state of the signal SDo. Therefore, the output signal (origin signal SZn') output from the AND gate LC2 is output at the same timing as the origin signal SZn. That is, in the first state, the origin signal SZn and the origin signal SZn' can be regarded as the same. In the first state, the time interval Tpx of the origin signal SZn' applied to the single irradiation pulse generator LC3 is less than the time Tpd. Therefore, the signal SDo from the single irradiation pulse generator LC3 is maintained at "0". In addition, even if there is an error in the angle ηj between the respective reflecting surfaces RP of the polygonal mirror PM, the fact that the time interval of the origin signal SZn' is less than the time Tpd will not change.

When the second state is reached, in which the scanning of the spot light SP skips over one reflection surface RP, the state signal STS switches to "0." Therefore, the output signal of the OR gate LC1 becomes "1" only when the signal SDo is "1." When the origin signal SZn (for ease of explanation, this origin signal SZn will be referred to as the first origin signal SZn) is applied while the signal SDo is "1" (in this case, the output signal of the OR gate LC1 is also "1"), the AND gate LC2 also outputs the origin signal SZn' in response thereto. However, when the origin signal SZn' is generated, the signal SDo from the single-shot pulse generator LC3 changes to "0" within the time Tpd. Therefore, during the time Tpd, since both inputs of the OR gate LC1 are "0" signals, the output signal of the OR gate LC1 remains "0." Therefore, during the time Tpd, the output signal of the AND gate LC2 also remains "0." Therefore, even if the second origin signal SZn is applied to the AND gate LC2 before the time Tpd elapses, the AND gate LC2 does not output the origin signal SZn'.

Then, after time Tpd has passed, signal SDo from single-shot pulse generator LC3 inverts to "1." Similar to the case of the first origin signal SZn, AND gate LC2 outputs origin signal SZn', corresponding to the third origin signal SZn applied after time Tpd. By repeating this operation, logic circuit LCC converts origin signal SZn, which is generated repeatedly with time Tpx, into origin signal SZn', which is generated repeatedly with time 2 × Tpx. From another perspective, logic circuit LCC generates origin signal SZn' by thinning out every other pulse of origin signal SZn, which is generated repeatedly with time Tpx. In other words, it divides the frequency of the generation timing of origin signal SZn by 1/2. Alternatively, logic circuit LCC of sub-origin generation circuit CAan can be replaced with frequency divider 330 ( FIG. 31 ) of sub-origin generation circuit CAn, as described in the fourth embodiment. When the frequency divider 330 is replaced, the frequency divider 330 only needs to divide the origin signal SZn into 1/2 in the second state and not divide the origin signal SZn in the first state. Alternatively, the sub-origin generation circuit CAn of the fourth embodiment can be replaced with the sub-origin generation circuit CAan of the fifth embodiment. Furthermore, in the second state, the origin signal SZ1' output from the logic circuit LCC of the sub-origin generation circuit CAa1 is offset in phase by half a cycle from the origin signal SZ4' output from the logic circuit LCC of the sub-origin generation circuit CAa4. Similarly, the origin signals SZ2' and SZ3' output from the logic circuit LCC of the sub-origin generation circuits CAa2 and CAa3 are offset in phase by half a cycle from the origin signals SZ5' and SZ6' output from the logic circuit LCC of the sub-origin generation circuits CAa5 and CAa6.

In this way, simply by inverting the value of the state signal STS input to the logic circuit LCC of each sub-origin generating circuit CAa1 to CAa6 of the light beam switching control unit 352, it is possible to arbitrarily switch between the first state in which the drawing exposure is repeatedly performed based on the scanning of the point light SP for each continuous reflecting surface RP of the polygonal mirror PM, and the second state in which the drawing exposure is repeatedly performed based on the scanning of the point light SP for each reflecting surface RP of the polygonal mirror PM.

Furthermore, in this fifth embodiment, the rotation control of the polygon mirror PM of each scanning unit Un (U1-U6) is also similar to that of the fourth embodiment described above. Specifically, the rotation of the polygon mirror PM of each scanning unit Un (U1-U6) is controlled so that the origin signals SZn (SZ1-SZ6) output from the origin sensors OPn of each scanning unit Un (U1-U6) have the relationship shown in FIG34 . Therefore, in the first state, in which the spot light SP is scanned on each reflection surface RP without skipping any surfaces, the scanning units U1-U3 can repeatedly scan the spot light SP in the order U1 → U2 → U3, and the scanning units U4-U6 can repeatedly scan the spot light SP in the order U4 → U5 → U6.

Preferably, the time Tpd set for the single-shot irradiation pulse generator LC3 can be changed based on information on the rotation speed of the polygon mirror PM from the exposure control unit 356. Furthermore, the situation is not limited to skipping one surface. Even when the spot light SP is scanned while skipping two surfaces, as long as the structure shown in FIG39 is used, this can be addressed simply by setting the time Tpd to the relationship of (n+1)×Tpx>Tdp>n×Tpx. Furthermore, n represents the number of skipped reflection surfaces RP. For example, when n is 2, the spot light SP is scanned every two reflection surfaces RP, and when n is 3, the spot light SP is scanned every three reflection surfaces RP.

Next, we will briefly describe the output control of the drawing bit string data Sdw by the drawing data output control unit 354 for the driver circuits 206a of the light source devices 14A' and 14B' during the first state, in which the spot light SP scans each reflective surface RP without skipping any surfaces. In the first state, the first scanning module (scanning units U1-U3) and the second scanning module (scanning units U4-U6) scan the spot light SP in parallel. Therefore, the drawing data output control unit 354 outputs the drawing bit string data Sdw, which is a time-sequential synthesis of the serial data DL1-DL3 corresponding to the scanning units U1-U3, to the driver circuit 206a of the light source device 14A' that emits the light beam LBa incident on the first scanning module. It also outputs the drawing bit string data Sdw, which is a time-sequential synthesis of the serial data DL4-DL6 corresponding to the scanning units U4-U6, to the driver circuit 206a of the light source device 14B' that emits the light beam LBb incident on the second scanning module.

Furthermore, the drawing data output control unit 354 shown in FIG35 can be used almost directly regardless of whether the state signal STS is "1" or "0." In the first state, in which the spot light SP scans each reflection surface RP without skipping surfaces, the sub-origin signal ZP2 is generated time Ts after the sub-origin signal ZP1 is generated, and the sub-origin signal ZP3 is generated time Ts after the sub-origin signal ZP1 is generated. Therefore, the serial data DL1 to DL3 are repeatedly output by the counter units CN1 to CN3 in the order of DL1 → DL2 → DL3. The serial data DL1 to DL3, which are sequentially output by the gate units GT1 to GT3, which are opened for a predetermined time (on time Ton) after the sub-origin signals ZP1 to ZP3 are applied, are input to the driver circuit 206a of the first light source device 14A' as the drawing bit string data Sdw. Similarly, in the first state, where spot light SP scans each reflection surface RP without skipping any surfaces, sub-origin signal ZP5 is generated time Ts after sub-origin signal ZP4 is generated, followed by sub-origin signal ZP6 time Ts after that. Consequently, serial data DL4 to DL6 are repeatedly output by counter sections CN4 to CN6 in the order DL4 → DL5 → DL6. This serial data DL4 to DL6, sequentially output by gate sections GT4 to GT6, which are opened for a predetermined time (on time Ton) after application of sub-origin signals ZP4 to ZP6, is input as drawing bit string data Sdw to the driver circuit 206a of the second light source device 14B'.

Next, we briefly explain the offsets of serial data DL1 to DL6 in the first state. Serial data DL1 is offset in the column direction at the timing when the sub-origin signal ZP2 corresponding to the scanning unit U2, which is scanning after the output of serial data DL1, is generated. Serial data DL2 is offset in the column direction at the timing when the sub-origin signal ZP3 corresponding to the scanning unit U3, which is scanning after the output of serial data DL2, is generated. Serial data DL3 is offset in the column direction at the timing when the sub-origin signal ZP1 corresponding to the scanning unit U1, which is scanning after the output of serial data DL3, is generated. Furthermore, serial data DL4 is offset in the column direction at the timing when the sub-origin signal ZP5 corresponding to the scanning unit U5, which is scanning after the output of serial data DL4, is generated. Serial data DL5 is offset in the column direction at the timing when the sub-origin signal ZP6 corresponding to the scanning unit U6, which is scanning after the output of serial data DL5, is generated. The offset in the column direction of the serial data DL6 is determined by the timing of the generation of the corresponding sub-origin signal ZP4 by the scanning unit U4, which then performs scanning after the output of the serial data DL6 is completed. Furthermore, the output control of the drawing bit string data Sdw in the second state is the same as in the fourth embodiment, so its description is omitted. Furthermore, the output control of the drawing bit string data Sdw in the first state follows the same control principle as in the first to third embodiments described above, with only the order of the output serial data DLn differing. Specifically, the difference lies in whether the serial data DLn is output in the order DL1 → DL3 → DL5, DL2 → DL4 → DL6, or in the order DL1 → DL2 → DL3, DL4 → DL5 → DL6.

In addition, in the second state where the scanning of the point light SP is performed by skipping one reflection surface RP, the scanning start interval of the point light SP of each scanning unit Un (U1 to U6) is longer than that in the first state where the scanning is performed for each reflection surface RP without skipping the surface. For example, in the case where the scanning of the point light SP is performed by skipping one reflection surface RP, the scanning start interval of the point light SP of each scanning unit Un (U1 to U6) is doubled compared to the case where the surface is not skipped. In addition, in the case where the scanning is performed by skipping two reflection surfaces RP, the scanning start interval of the point light SP is tripled compared to the case where the surface is not skipped. Therefore, if the rotation speed of the polygon mirror PM and the transport speed of the substrate FS are made the same in the first state and the second state, the exposure results become different in the first state and the second state.

Therefore, the exposure control unit 356 may also have a control mode for changing (correcting) at least one of the rotation speed of the polygon mirror PM and the conveying speed of the substrate FS in the first state and the second state so that the exposure results in the first state and the second state are the same. For example, when the scanning start interval of the point light SP in the first state and the scanning start interval of the point light SP in the second state are 1:2, the exposure control unit 356 controls the rotation control unit 350 so that the ratio of the rotation speed of the polygon mirror PM in the first state to the rotation speed of the polygon mirror PM in the second state becomes 1:2. Specifically, the rotation speed of the polygon mirror PM in the first state is 20,000 rpm, and the rotation speed of the polygon mirror PM in the second state is 40,000 rpm. Accordingly, for example, if the light emission frequency Fs of the light beam LB (LBa, LBb) of the light source device 14' (14A', 14B') is 200 MHz in the first state, it is set to 400 MHz in the second state. Thus, the interval of the generation timing of the sub-origin signal ZPn in the first state and the interval of the generation timing of the sub-origin signal ZPn in the second state can be made substantially the same.

In addition, for example, the exposure control unit 356 may also have a control mode for controlling the rotation speed of the drive rollers R1 to R3 and the rotating drum DR in a manner such that the ratio of the conveying speed of the substrate FS in the first state to the conveying speed of the substrate FS in the second state becomes 2:1 when the scanning start interval of the point light SP in the first state and the scanning start interval of the point light SP in the second state are 1:2. By using either the control mode (scanning correction mode) for correcting the rotation speed of the polygon mirror PM and/or the light emission frequency Fs (frequency of the clock signal LTC) and the control mode (conveying correction mode) for correcting the conveying speed of the substrate FS, the interval in the X direction of the drawing lines SLn (SL1 to SL6) on the substrate FS in the first state and the interval in the X direction of the drawing lines SLn (SL1 to SL6) on the substrate FS in the second state can be made the same interval (for example, 1.5 μm). Furthermore, in the first state and the second state, the pattern data (bitmaps) stored in the memory units BM1 to BM6 in the drawing data output control unit 354 can be used as they are without any modification.

In addition, both the scanning correction mode and the transport correction mode described above can be used to perform corrections in a manner that the pattern drawn on the substrate FS in the first state is the same as the pattern drawn on the substrate FS in the second state. For example, in the first state (the case of light beam scanning according to each reflection surface RP of the polygon mirror PM), when the rotation speed of the polygon mirror PM is 20,000 rpm, the luminous frequency Fs of the light beam LB of the light source device 14' (14A', 14B') is 200 MHz, and the transport speed of the substrate FS is 5 mm/second, in the second state (the case of light beam scanning that skips one reflection surface RP of the polygon mirror PM), the transport speed of the substrate FS can be set to 3.75 mm/second instead of being slowed down by half, and the rotation speed of the polygon mirror PM can be set to 1.5 times that of 30,000 rpm, and the luminous frequency Fs of the light beam LB can also be set to 1.5 times that of 300 MHz. In this way, if the scanning correction mode and the transport correction mode are combined, in the case of the second state, there is no need to reduce the transport speed of the substrate FS to half, thereby suppressing an extreme decrease in productivity.

Furthermore, in the fifth embodiment, as described in the fourth embodiment, the number of scanning units Un that distribute the light beams LBa and LBb can be arbitrarily changed. Furthermore, the scanning efficiency of the polygonal mirror PM can also be arbitrarily changed. Furthermore, in the fifth embodiment, since the scanning efficiency of the polygonal mirror PM is set to 1/3 and the number of scanning units Un is set to six, the six selection optical elements AOMn (AOM1 to AOM6) are divided into two optical element modules OM1 and OM2, and the six scanning units Un (U1 to U6) are correspondingly divided into two scanning modules. However, in the case where the scanning efficiency of the polygonal mirror PM is 1/M and the number of scanning units Un and selection optical elements AOMn is Q, it is sufficient to divide the Q selection optical elements AOMn into Q/M optical element modules OM1, OM2, ..., and divide the Q scanning units Un into Q/M scanning modules. In this case, it is preferred that each optical element module OM1, OM2, ..., include an equal number of selection optical elements AOMn. Furthermore, the number of scanning units Un included in each of the Q/M scanning modules is also equal. Furthermore, Q/M is preferably an integer. In other words, Q is preferably a multiple of M.

For example, when the scanning efficiency of the polygon mirror PM is 1/2 and the number of scanning units Un and selection optical elements AOMn is six, the six selection optical elements AOMn can be equally divided into three optical element modules OM1, OM2, and OM3, and the six scanning units Un can be equally divided into three scanning modules. Furthermore, in the first state, the three optical element modules OM1, OM2, and OM3 can be arranged in parallel, and the light beams LB (in this case, LBa, LBb, and LBc) from the three light source devices 14' can be incident on each of the three optical element modules OM1, OM2, and OM3 in parallel. In the second state, the three optical element modules OM1, OM2, and OM3 can be arranged in series, and the light beam LB from the one light source device 14' can be incident on the three optical element modules OM1, OM2, and OM3 in sequence.

As described above, in the fifth embodiment, the beam switching control unit 352 controls the beam switching component 20A to sequentially perform one-dimensional scanning of the spot light SP by each of the plurality of scanning units Un, such that the deflection (scanning) of the light beam LBn (spot light SP) by the polygon mirror PM of the scanning unit Un is switched between a first state (first drawing mode) in which the deflection is repeated for each successive reflection surface RP of the polygon mirror PM, and a second state (second drawing mode) in which the deflection is repeated for every at least one reflection surface RP of the polygon mirror PM. This achieves the same effects as those of the fourth embodiment, and allows switching between scanning the spot light SP with or without skipping a surface.

In the first state, when the scanning efficiency (α/β) of the polygon mirror PM is less than 1/2, a number of scanning units Un corresponding to the inverse of the scanning efficiency are grouped into a scanning module. Using multiple scanning modules in this group, one scanning unit Un is configured for each scanning module to perform one-dimensional scanning of the point light SP. This allows the point light SP to simultaneously scan a number of the same drawing lines SLn as the number of scanning modules. Furthermore, in the second state, since the light beam is controlled to scan at least every other reflective surface RP of the polygon mirror PM, even with a greater number of scanning units Un than the inverse of the scanning efficiency (α/β) of the polygon mirror PM, the light beam LB can be effectively utilized, and all of these scanning units Un scan the point light SP along the drawing lines SLn.

In the case of the first state mentioned above, since the light beams LBa and LBb from the light source devices 14A’ and 14B’ are incident on the two grouped scanning modules in parallel, the selection optical elements AOM1~AOM6 in the light beam switching component 20A are respectively switched on/off through the light beam switching control unit 352, in units of the grouped scanning modules, in a time-sharing manner so as to be incident on the scanning units U1~U6 corresponding to the light beams LB1~LB6.

The configuration switching component SWE provided in the beam switching component 20A switches between a first configuration state and a second configuration state. In the first configuration state, the three selection optical elements AOM1 to AOM3 are connected in series along the optical path of the beam LBa, and the selection optical elements AOM4 to AOM6 are connected in series along the optical path of the beam LBb in such a manner that the beam LBa from the first light source device 14A' is distributed as beams LB1 to LB3 to three scanning units U1 to U3 among the six scanning units U1 to U6, and the beam LBb from the second light source device 14B' is distributed as beams LB4 to LB6 to the remaining three scanning units U4 to U6. In the second configuration state, the six selection optical elements AOM1 to AOM6 are connected in series along the optical path of the beam LBa in such a manner that the beam LBa from one light source device 14A' is distributed as beams LB1 to LB6 to the six scanning units U1 to U6.

Thus, in the first state, the switching element SWE is configured to set the first configuration state. As a result, each scanning unit U1-U6 can repeatedly scan with the point light SP along each successive reflection surface RP of the polygon mirror PM, and two of the six scanning units U1-U6 can scan with the point light SP almost simultaneously. Furthermore, in the second state, the switching element SWE is configured to set the second configuration state. Thus, although light beam scanning is performed along at least every other reflection surface RP of the polygon mirror PM, scanning with the point light SP can be repeatedly performed by all six scanning units U1-U6.

Therefore, according to this fifth embodiment, during the initial installation of the drawing device, when a light source device 14A' is used to set the configuration switching component SWE in a manner that is in the second configuration state, and then when it is desired to increase the conveying speed of the substrate FS, it is sufficient to add a second light source device 14B' and set the configuration switching component SWE in a manner that is in the first configuration state. In terms of hardware, the drawing device can be upgraded through simple operations such as adding a light source device and switching the configuration switching component SWE.

Furthermore, in each of the above-described embodiments, the origin signal SZn is detected using the reflection surface RP located immediately before the reflection surface RP of the polygon mirror PM in the rotational direction relative to the reflection surface RP that deflects the light beam LBn. However, the origin signal SZn may also be detected using the reflection surface RP itself that deflects the light beam LBn. In this case, the origin signal SZn or the origin signal SZn′ derived from the origin signal SZn does not need to be delayed by the time Tpx. Therefore, the origin signal SZn or the origin signal SZn′ can be simply used as the sub-origin signal ZPn.

In the fourth and fifth embodiments, the electro-optical element 206 serving as the drawing light modulator of the light source device 14' (14A', 14B') is switched using the drawing bit string data Sdw. However, as in the second embodiment, a drawing optical element AOM may be used as the drawing light modulator. This drawing optical element AOM is an acousto-optic modulator (AOM). That is, in the fourth embodiment, the drawing optical element AOM may be positioned between the light source device 14' and the primary selection optical element AOM1, so that the light beam LB from the light source device 14' that has passed through the drawing optical element AOM is incident on the selection optical element AOM1. In this case, the drawing optical element AOM is switched based on the drawing bit string data Sdw. Even in this case, the same effects as those of the fourth embodiment can be achieved.

Furthermore, in the fifth embodiment described above, image drawing optical elements AOM (AOMa, AOMb) are disposed between the first light source device 14A' and the primary selection optical element AOM1 of the first optical element module OM1, and between the second light source device 14B' and the primary selection optical element AOM4 of the second optical element module OM2. Specifically, the light beam LBa from the light source device 14A' that has passed through the image drawing optical element AOMa enters the selection optical element AOM1, and the light beam LBb from the light source device 14B' that has passed through the image drawing optical element AOMb enters the selection optical element AOM4. In this case, in the first state, the image drawing optical element AOMa switches according to the image drawing bit string data Sdw consisting of the serial data DL1 to DL3, and the image drawing optical element AOMb switches according to the image drawing bit string data Sdw consisting of the serial data DL4 to DL6. In the second state, only the image drawing optical element AOMa is switched based on the image drawing bit string data Sdw composed of the serial data DL1 to DL6.

Alternatively, as in the first embodiment, an image drawing optical element AOM serving as an image drawing light modulator may be provided for each scanning unit Un. In this case, the image drawing optical element AOM may be provided just before the reflector M20 (see FIG. 28 ) of each scanning unit Un. The image drawing optical element AOM of each scanning unit Un (U1 to U6) is switched based on the serial data DLn (DL1 to DL6). For example, the image drawing optical element AOM of scanning unit U3 is switched based on the serial data DL3.

[Sixth embodiment]

FIG41 illustrates the structure of a beam switching component (beam delivery unit) 20B according to a sixth embodiment. Here, the light beam LBw (LB) emitted from a light source device 14' and incident on the beam switching component 20B is assumed to be a circularly polarized parallel beam. The beam switching component 20B includes six selection optical elements AOM1 to AOM6, two absorbers TR1 and TR2, six lens systems CG1 to CG6, mirrors M30, M31, and M32, a focusing lens CG0, a polarization beam splitter BS1, and two image forming optical elements (acousto-optic modulators) AOMa and AOMb. Components identical to those in the fourth or fifth embodiment are denoted by the same reference numerals.

The light beam LBw incident on the beam switching component 20B passes through the focusing lens CG0 and is separated by the polarization beam splitter BS1 into a linear P-polarized light beam LBp and a linear S-polarized light beam LBs. The S-polarized light beam LBs, reflected by the polarization beam splitter BS1, is incident on the image drawing optical element AOMa. The light beam LBs incident on the image drawing optical element AOMa is converged by the focusing action of the focusing lens CG0 to form a beam waist within the image drawing optical element AOMa. The image drawing bit string data Sdw (DLn) shown in Figure 19 is applied to the image drawing optical element AOMa via the driver circuit DRVn. This image drawing bit string data Sdw is synthesized by combining the serial data DL1, DL3, and DL5 corresponding to the odd-numbered scanning units U1, U3, and U5, respectively. Therefore, when the depiction bit string data Sdw (DLn) is "1", the depiction optical element AOMa is in the On state, and the first-order diffracted light of the incident light beam LBs is emitted toward the mirror M31 as a deflected depiction light beam (an intensity-modulated light beam). The depiction light beam reflected by the mirror M31 passes through the lens system CG1 and is incident on the selection optical element AOM1. In addition, when the depiction bit string data Sdw (DLn) is "0", the zero-order light (LBs) emitted from the depiction optical element AOMa is reflected by the mirror M31, but travels at an angle that does not enter the subsequent lens system CG1. In addition, the lens system CG1 converges the depiction light beam divergently emitted from the depiction optical element AOMa at the diffraction part of the selection optical element AOM1 to form a beam waist.

The drawing light beam transmitted through the selection optical element AOM1 enters the selection optical element AOM3 via lens system CG3, which is identical to lens system CG1. The drawing light beam transmitted through the selection optical element AOM3 enters the selection optical element AOM5 via lens system CG5, which is identical to lens system CG1. FIG41 illustrates a state in which the three selection optical elements AOM1, AOM3, and AOM5 are arranged in series along the beam path, with only the selection optical element AOM3 in the ON state. The drawing light beam, intensity-modulated by the drawing optical element AOMa, enters the corresponding scanning unit U3 as light beam LB3. Furthermore, the lens systems CG1, CG3, and CG5 are equivalent to a combination of a collimating lens CL and a focusing lens CD as shown in FIG26 or FIG36.

On the other hand, the P-polarized light beam LBp transmitted through the polarization beam splitter BS1 is reflected by the mirror M30 and incident on the drawing optical element AOMb. The light beam LBp incident on the drawing optical element AOMb is converged by the focusing effect of the focusing lens CG0 to form a beam waist in the drawing optical element AOMb. In the drawing optical element AOMb, the drawing bit string data Sdw (DLn) shown in Figure 19 is applied via the driver circuit DRVn. The drawing bit string data Sdw is obtained by synthesizing the serial data DL2, DL4, and DL6 corresponding to the even-numbered scanning units U2, U4, and U6. Therefore, when the drawing bit string data Sdw (DLn) is "1", the drawing optical element AOMb is in the On state, and the first diffracted light of the incident light beam LBp is emitted toward the mirror M32 as a deflected drawing light beam (an intensity-modulated light beam). The drawing light beam reflected by mirror M32 passes through lens system CG2, the same lens system as lens system CG1, and enters the selection optical element AOM2. Furthermore, when the drawing bit string data Sdw (DLn) is "0," the zero-order light (LBp) emitted from the drawing optical element AOMb is reflected by mirror M32, but travels at an angle that prevents it from entering the subsequent lens system CG2. Furthermore, lens system CG2 converges the drawing light beam, which diverges from the drawing optical element AOMb, at the diffracted portion of the selection optical element AOM2, forming a beam waist.

The drawing light beam transmitted through the selection optical element AOM2 enters the selection optical element AOM4 via lens system CG4, which is identical to lens system CG1. The drawing light beam transmitted through the selection optical element AOM4 enters the selection optical element AOM6 via lens system CG6, which is identical to lens system CG1. Figure 41 shows the following state: the three selection optical elements AOM2, AOM4, and AOM6 are arranged in series along the beam optical path, with only the selection optical element AOM2 in the on state. The drawing light beam, intensity-modulated by the drawing optical element AOMb, enters the corresponding scanning unit U2 as light beam LB2. Furthermore, the lens systems CG2, CG4, and CG6 are equivalent to a combination of a collimating lens CL and a focusing lens CD as shown in Figures 26 or 36.

If a beam switching component (beam distribution unit) 20B as shown in Figure 41 above is used, the light beam LBw from a light source device 14' can be split into two beams by the polarization beam splitter BS1, and the light beam LBs from one side is passed through the drawing optical element AOMa to generate a drawing light beam (LB1, LB3, LB5) that is incident on one of the odd-numbered scanning units U1, U3, U5 in sequence, and the light beam LBp from the other side split by the polarization beam splitter BS1 is passed through the drawing optical element AOMb to generate a drawing light beam (LB2, LB4, LB6) that is incident on one of the even-numbered scanning units U2, U4, U6 in sequence.

In this sixth embodiment, after the light beam LBw from the light source device 14' is split into two beams by the polarization beam splitter BS1, the intensity of the light beam LB is modulated based on the pattern data by the drawing optical elements AOMa and AOMb. Therefore, if the intensity of the spot light SP of each of the six scanning units U1 to U6 is attenuated to -50% in the polarization beam splitter BS1, -20% in the drawing optical elements AOMa and AOMb, and -30% within each scanning unit U1 to U6, the intensity is approximately 22.4% of the original intensity of the light beam LBw (100%). However, if the scanning efficiency of the polygon mirror PM of each of the six scanning units U1 to U6 is less than 1/3 and the light beam LBw from a single light source device 14' is used, the beam scanning is performed without skipping over a single reflective surface RP of the polygon mirror PM, and pattern drawing can be performed using the scanning of the spot light SP on each of the six drawing lines SLn.

[Variant 1]

As in the sixth embodiment, when the polarization directions of the light beams LBs incident on the odd-numbered selective optical elements AOM1, AOM3, and AOM5 are orthogonal to the polarization directions of the light beams LBp incident on the even-numbered selective optical elements AOM2, AOM4, and AOM6, the odd-numbered selective optical elements AOMn and the even-numbered selective optical elements AOMn need to be arranged so as to be rotated 90 degrees relative to each other about the beam incidence axis. Figure 42 shows, for example, a configuration in which, among the odd-numbered selective optical elements AOM1, AOM3, and AOM5, selective optical element AOM3 is arranged so as to be rotated 90 degrees relative to the even-numbered selective optical element AOMn. Selective optical element AOM3 receives the S-polarized drawing beam that has passed through lens system CG3, so the direction of high diffraction efficiency is the Y direction, which is parallel to the XY plane. In other words, selective optical element AOM3 is arranged so as to rotate 90 degrees so that the periodic direction of the diffraction grating generated within selective optical element AOM3 is in the Y direction.

With this configuration of the selection optical element AOM3, the deflected light beam LB3 emitted when the selection optical element AOM3 is in the "on" state travels obliquely in the Y direction relative to the direction of travel of the zero-order light. Therefore, a mirror IM3a is provided to separate the light beam LB3 from the optical path of the zero-order light and reflect the light beam LB3 from the selection optical element AOM3 within the XY plane, allowing the light beam LB3 to pass through the opening TH3 of the support member IUB in the Z direction. A mirror IM3b is also provided to reflect the light beam LB3 reflected by the mirror IM3a in the -Z direction, allowing the light beam LB3 to pass through the opening TH3. Similarly, a pair of mirrors IM1a and IM1b and a pair of mirrors IM5a and IM5b are provided for the other odd-numbered selection optical elements AOM1 and AOM5. Moreover, in the structure of Figure 41, since the polarization directions of the light beams LBs and LBp incident on the drawing optical elements AOMa and AOMb are orthogonal, the drawing optical elements AOMa and AOMb are arranged in a relationship of being relatively rotated 90 degrees around the beam incident axis.

However, when the polarization beam splitter BS1 in FIG41 is an amplitude-splitting beam splitter or a half-mirror, if the polarization direction of the light beam LBw is set to only one direction (for example, P polarization), there is no need to rotate one of the depicting optical elements AOMa, AOMb, the odd-numbered selection optical element AOMn, and one of the even-numbered selection optical elements AOMn relative to each other by 90 degrees as shown in FIG42.

[Variant 2]

In the sixth embodiment, the scanning units U1 to U6 corresponding to the six selective optical elements AOM1 to AOM6 are all capable of scanning the point light SP along the respective drawing lines SL1 to SL6 according to all the reflection surfaces RP of the polygon mirror PM. Therefore, in order to receive the light beams (light beams modulated by the drawing optical element AOMa) that pass through the odd-numbered selective optical elements AOM1, AOM3, and AOM5 in sequence, three selective optical elements AOM7, AOM9, and AOM11 are further arranged in series between the selective optical element AOM5 and the absorber TR2 in FIG. 41 , and in order to receive the light beams (light beams modulated by the drawing optical element AOMb) that pass through the even-numbered selective optical elements AOM2, AOM4, and AOM6 in sequence, three selective optical elements AOM8, AOM10, and AOM12 are further arranged in series between the selective optical element AOM6 and the absorber TR1. Furthermore, six scanning units U7 to U12 are added to guide the light beams LB7 to LB12 deflected (switched) by the selective optical elements AOM7 to AOM12, respectively. This results in a total of 12 scanning units U1 to U12 arranged across the width (Y) direction of the substrate FS. This enables combined drawing exposure for 12 drawing lines SL1 to SL12, doubling the maximum exposure width in the Y direction.

In this case, when the scanning efficiency of the polygon mirror PM of each of the scanning units U1-U12 is less than 1/3, the odd-numbered scanning units U1, U3, U5, U7, U9, and U11 grouped as the first drawing module and the even-numbered scanning units U2, U4, U6, U8, U10, and U12 grouped as the second drawing module all scan the light beam LBn across every reflection surface RP of the polygon mirror PM. In this way, even if the Y-direction width of the substrate FS increases, pattern drawing can be performed on a larger exposure area W (Figures 5 and 25) simply by adding scanning units U7-U12 and selecting optical elements AOM7-AOM12. In this way, the structure of adding six scanning units U7~U12 and selecting optical elements AOM7~AOM12 to become 12 scanning units U1~U12 can be similarly applied to the situation of using two light source devices 14A', 14B' described in the previous fifth embodiment (Figures 36 to 38).

[Variant 3]

FIG43 shows the relationship between the transport configuration of the substrate FS and the arrangement of the scanning units Un (drawing lines SLn) in Modification 3. Here, as in Modification 2, 12 scanning units U1 to U12 are provided and arranged on the rotating drum DR so that the drawing lines SL1 to SL12 of each scanning unit Un can be joined and drawn and exposed in the Y direction. Furthermore, the length of the rotating drum DR and the various rollers R1 to R3, RT1, RT2, etc. in the substrate transport mechanism 12 shown in FIG23 in the direction of the rotation axis (Y direction) is denoted as Hd, the maximum scanning width in the Y direction that can be exposed by the joined drawing of the 12 scanning units Un is denoted as Sh (Sh < Hd), and the maximum support width of the substrate FS0 that can be exposed is denoted as Tf. In Modification 3, the twelve scanning units U1-U12 corresponding to the twelve drawing lines SL1-SL12 are each configured to receive the twelve corresponding light beams LB1-LB12 in a time-division manner from a beam switching component (beam distribution unit) 20B that splits the light beam LBw from a single light source device 14' into two using a beam splitter and a half-mirror, as shown in FIG41 (Sixth Embodiment), or from a beam switching component (beam distribution unit) 20A that uses the light beams LBa and LBb from two light source devices 14A' and 14B', as shown in FIG38 (Fifth Embodiment). Therefore, for example, if the length of each drawing line SL1-SL12 in the Y direction is 50 mm, the maximum scanning width Sh is 600 mm. As an example, the width of the substrate FS0, which constitutes the maximum support width Tf, can be 650 mm, and the length Hd of the rotating drum DR can be approximately 700 mm.

When exposing a substrate FS0 having the same width as the maximum support width Tf using a drawing apparatus such as that shown in FIG43 , in addition to the four alignment microscopes AM1 to AM4 (observation areas Vw1 to Vw4) shown in FIG24 and FIG25 , three additional alignment microscopes AM5 to AM7 (observation areas Vw5 to Vw7) are provided in the Y direction. In this case, alignment microscope AM1 (observation area Vw1) and alignment microscope AM7 (observation area Vw7), located on either side of the width of the substrate FS0, detect alignment marks formed at regular intervals along the X direction on both sides of the substrate FS0. Furthermore, alignment microscope AM4 (observation area Vw4) is positioned approximately in the center of the maximum support width Tf.

In addition, in the case of a substrate FS1 on which a pattern can be drawn on the exposure area W by the respective drawing lines SL1 to SL6 of the six scanning units U1 to U6 as described in the previous embodiments, its width Tf1 is approximately half of the maximum support width Tf of the rotating drum DR. Therefore, the substrate FS1 is transported, for example, on the -Y direction side of the outer peripheral surface of the rotating drum DR. At this time, the alignment marks MK1 to MK4 (Figure 25) on the substrate FS1 can be detected by the observation areas Vw1 to Vw4 of the four alignment microscopes AM1 to AM4. Moreover, in the case of exposing the substrate FS1, since only six scanning units U1 to U6 are used, the scanning units U1 to U6 can each perform point scanning along each drawing line SL1 to SL6, whether in a mode of beam scanning at each continuous reflection surface RP of the polygon mirror PM or in a mode of beam scanning at every reflection surface RP of the polygon mirror PM.

For example, when it is set to use the light beams LBa and LBb from the two light source devices 14A’ and 14B’ at the same time as in the fifth embodiment, the light beam LBa from the light source device 14A’ is grouped in the light beam switching component 20A in a manner that the light beam LBa from the light source device 14A’ is transmitted in series from the selection optical elements AOM1, AOM3, AOM5, AOM7, AOM9, AOM11 corresponding to the odd-numbered scanning units U1, U3, U5, U7, U9, U11, and the light beam LBa from the light source device 14A’ is grouped in the light beam switching component 20A in a manner that the light beam LBa from the light source device 14A’ is transmitted in series from the selection optical elements AOM2, AOM4, AOM6, AOM8, AOM10, AOM12 corresponding to the even-numbered scanning units U2, U4, U6, U8, U10, U12. Moreover, during the exposure of the substrate FS1, the exposure is controlled in a manner that light beam scanning is repeatedly performed on each continuous reflection surface RP of the polygonal mirror PM in the order of odd-numbered scanning units U1, U3, U5, based only on the three origin signals SZ1, SZ3, SZ5 output from each continuous reflection surface RP of the polygonal mirror PM; and the exposure is controlled in a manner that light beam scanning is repeatedly performed on each continuous reflection surface RP of the polygonal mirror PM in the order of even-numbered scanning units U2, U4, U6, based only on the three origin signals SZ2, SZ4, SZ6 output from each continuous reflection surface RP of the polygonal mirror PM.

Furthermore, when exposing a substrate FS2 having a width Tf2 smaller than the maximum support width Tf and larger than the width Tf1 of the substrate FS1, the substrate FS2 is conveyed so as to align with the center portion of the maximum support width Tf of the rotating drum DR. At this time, the exposure area W on the substrate FS2 can be drawn by the drawing lines SL3 to SL10 of the eight scanning units U3 to U10 connected in the Y direction. In this case, control is performed so that the four odd-numbered selection optical elements AOM3, AOM5, AOM7, and AOM9, which receive the light beam LBa (intensity-modulated light beam) from the light source device 14A', sequentially generate one of the light beams LB3, LB5, LB7, and LB9 in time-sharing order, and the four even-numbered selection optical elements AOM4, AOM6, AOM8, and AOM10, which receive the light beam LBb (intensity-modulated light beam) from the light source device 14B', sequentially generate one of the light beams LB4, LB6, LB8, and LB10 in time-sharing order. Therefore, each of the at least eight scanning units U3 to U10 is set in a mode of performing light beam scanning on every reflecting surface RP of the polygon mirror PM.

Moreover, during the exposure of the substrate FS2, the exposure is controlled in a manner that the light beam scanning is repeatedly performed on every reflection surface RP of the polygon mirror PM in the order of the odd-numbered scanning units U3, U5, U7, U9 based on the four sub-origin signals ZP3, ZP5, ZP7, and ZP9 outputted from every reflection surface RP of the polygon mirror PM only by the odd-numbered scanning units U3, U5, U7, and U9; and the exposure is controlled in a manner that the light beam scanning is repeatedly performed on every reflection surface RP of the polygon mirror PM in the order of the even-numbered scanning units U4, U6, U8, and U10 based on the four sub-origin signals ZP4, ZP6, ZP8, and ZP10 outputted from every reflection surface RP of the polygon mirror PM only by the even-numbered scanning units U4, U6, U8, and U10. In addition, in FIG43 , the alignment marks formed on both sides of the substrate FS2 in the width direction (equivalent to the alignment marks MK1 and MK4 in FIG25 ) are arranged so as to be detectable in the observation areas Vw2 and Vw6 of the alignment microscopes AM2 and AM6, respectively. However, depending on the size of the exposure area W in the Y direction, there is a case where it is not necessary to arrange them in such a relationship. In this case, it is sufficient to configure some of the seven alignment microscopes AM1 to AM7 to be movable in the Y direction so that the positional spacing of the observation areas Vw1 to Vw7 in the Y direction can be adjusted.

According to the above modification example 3, it is possible to perform highly efficient exposure using only the necessary scanning units Un according to the width of the substrate FS to be exposed and/or the Y-direction size of the exposure area W. In addition, when the scanning efficiency of the polygon mirror PM of each of the 12 scanning units U1 to U12 is 1/3 or less as shown in FIG43 , for example, as long as the light beam is scanned every three reflecting surfaces RP of each polygon mirror PM, even a light beam from a single light source device 14' can well depict a pattern within the range of the maximum scanning width Sh.

Furthermore, when the drawing device is composed of nine scanning units U1 to U9, the five odd-numbered scanning units U1, U3, U5, U7, and U9 and the four even-numbered scanning units U2, U4, U6, and U8 are used. Therefore, when the pattern is drawn on the exposure area W along all the drawing lines SL1 to SL9 by the nine scanning units U1 to U9, if the scanning efficiency of the polygon mirror PM is 1/3 or less, for example, it is sufficient to perform light beam scanning on each reflecting surface RP of each polygon mirror PM. However, in this case, it is sufficient to repeatedly scan the points on the odd-numbered drawing lines SL1, SL3, SL5, SL7, SL9 by referring only to the sub-origin signals ZP1, ZP3, ZP5, ZP7, ZP9 generated from the respective origin signals SZn of the odd-numbered scanning units U1, U3, U5, U7, U9, and repeatedly scan the points on the even-numbered drawing lines SL2, SL4, SL6, SL8 by referring only to the sub-origin signals ZP2, ZP4, ZP6, ZP8 generated from the respective origin signals SZn of the even-numbered scanning units U2, U4, U6, U8.

As described above, in variant example 3, a pattern drawing method is provided, using a drawing device, in which a plurality of scanning units Un that scan the point light SP of the light beam from the light source device 14' along the drawing line SLn are configured so that the pattern drawn by each drawing line SLn is joined along the direction of the drawing line SLn (main scanning direction) on the substrate FS, and the plurality of scanning units and the substrate FS are moved relative to each other in a sub-scanning direction intersecting the main scanning direction. The pattern drawing method includes: selecting a specific scanning unit among the plurality of scanning units Un that corresponds to the width of the substrate FS in the main scanning direction, or the width of the exposure area on the substrate FS to be patterned in the main scanning direction, or the position of the exposure area; and selectively supplying a light beam that has been intensity-modulated based on the pattern data to be drawn by each specific scanning unit to each of the specific scanning units in sequence via a light beam distribution unit that distributes the light beam from the light source device 14'. Thus, in Modification 3, even if the width of the substrate FS changes or the width or position of the exposure area W on the substrate FS changes, precise pattern drawing with high bonding accuracy can be performed by appropriately determining the transport position of the substrate FS in the Y direction. In addition, in this case, the rotation speed and rotation angle phase can be synchronized not between all the polygon mirrors PM of the multiple scanning units, but only between the polygon mirrors PM of the specific scanning units that contribute to pattern drawing.

[Variant 4]

Furthermore, as another configuration of a drawing device using nine scanning units U1-U9, the drawing device can be divided into two groups, not by odd-numbered and even-numbered groups, but simply by the order in which the scanning units Un are arranged. Specifically, the drawing device can be divided into a first scanning module composed of six scanning units U1-U6 and a second scanning module composed of three scanning units U7-U9, with the first scanning module supplied with light beam LBa from the first light source device 14A' and the second scanning module supplied with light beam LBb from the second light source device 14B'. In this case, if the scanning efficiency (α/β) of the polygon mirror PM satisfies 1/4 < (α/β) ≤ 1/3, each of the six scanning units U1-U6 within the first scanning module scans the spot light SP along each drawing line SL1-SL6 by scanning the beam every time one of the reflecting surfaces RP of the polygon mirror PM, similar to the fourth embodiment ( FIG. 33 ).

In contrast, the three scanning units U7 to U9 in the second scanning module can each perform light beam scanning on each of the reflection surfaces RP of the polygon mirror PM. Therefore, if the three scanning units U7 to U9 each directly perform light beam scanning on each of the reflection surfaces RP of the polygon mirror PM, the repetitive time interval ΔTc1 of the scanning of the point light SP on each of the drawing lines SL1 to SL6 of the six scanning units U1 to U6 and the repetitive time interval ΔTc2 of the scanning of the point light SP on each of the drawing lines SL7 to SL9 of the three scanning units U7 to U9 will be in the relationship of ΔTc1 = 2ΔTc2. The pattern drawn on the substrate FS by the drawing lines SL1 to SL6 and the pattern drawn on the substrate FS by the drawing lines SL7 to SL9 will become different, and good joint exposure cannot be performed.

Therefore, each of the three scanning units U7-U9, which can perform light beam scanning on each of the reflective surfaces RP of the polygon mirror PM, is controlled so as to perform light beam scanning on each reflective surface RP of the polygon mirror PM. This control can be achieved by inputting the origin signals SZ7-SZ9 generated by the scanning units U7-U9 into the circuit of FIG31 or the sub-origin generating circuit CAan of FIG38 to generate sub-origin signals ZP7-ZP9. In response to the sub-origin signals ZP7-ZP9, the corresponding selection optical elements AOM7-AOM9 are sequentially turned on for a predetermined time Ton, and the drawing serial data DL7-DL9 corresponding to the patterns to be drawn by the drawing lines SL7-SL9 are sequentially transmitted to the driving circuit 206a of the electro-optical element 206 in the second light source device 14B'.

[Variant 5]

FIG44 illustrates the structure of a driver circuit DRVn for the selective optical element AOMn of Modification 5. As described in the previous embodiments and modifications, when the plurality of scanning units Un each perform light beam scanning on one or more reflective surfaces RP of the polygonal mirror PM, the light beams LB (LBa, LBb) emitted from the light source device 14' (14A', 14B') and the light beams LBs and LBp emitted from the drawing optical elements AOMa and AOMb pass through the plurality of selective optical elements AOMn arranged along their optical paths. In FIG44 , after passing through the selective optical elements AOM1 and AOM2, the light beam LB is switched by the selective optical element AOM3, generating a light beam LB3 directed toward the scanning unit U3. Generally, the optical material within the selective optical element AOMn has a relatively high transmittance for the light beam LB in the ultraviolet band (e.g., a wavelength of 355 nm), but has an attenuation rate of several percent.

When the transmittance of each selective optical element AOMn is set to 95%, when the selective optical element AOM3 is in the On state as shown in Figure 44, the intensity of the light beam LB incident on the selective optical element AOM3 is attenuated by the two selective optical elements AOM1 and AOM2, and thus becomes approximately 90% ( ^0.952 ) of the original light beam intensity (100%) incident on the selective optical element AOM1. Furthermore, when the six selective optical elements AOM1 to AOM6 are connected, the intensity of the light beam LB incident on the last selective optical element AOM6 is attenuated by the five selective optical elements AOM1 to AOM5, and thus becomes approximately 77% ( ^0.955 ) of the original light beam intensity (100%).

As a result, the intensities of the light beams LB incident on each of the six selective optical elements AOM1 to AOM6 are, in order, 100%, 95%, 90%, 85%, 81%, and 77%. This means that the intensities of the light beams LB1 to LB6 deflected and emitted by the selective optical elements AOM1 to AOM6 also change gradually at these ratios. Therefore, in this fifth variation, the driver circuits DRVn for each of the multiple selective optical elements AOMn shown in FIG38 adjust the drive conditions for the selective optical elements AOM1 to AOM6 to minimize fluctuations in the intensities of the light beams LB1 to LB6.

In FIG44 , since driver circuits DRV1 through DRV6 (DRV5 and DRV6 are omitted from illustration) all have the same structure, only driver circuit DRV1 will be described in detail. As previously shown in FIG38 , each driver circuit DRV1 through DRV6 receives inputs for setting the on-time Ton of each of the selective optical elements AOM1 through AOM6 (AOM5 and AOM6 are omitted from illustration in FIG44 ) and sub-origin signals ZP1 through ZP6. Furthermore, the structure of FIG44 also shares a high-frequency transmission source 400 for applying ultrasonic waves to each of the selective optical elements AOM1 through AOM6. The driver circuit DRV1 comprises: a switching element 401 that receives a high-frequency signal from a high-frequency transmission source 400 and switches at high speed whether to transmit the signal to an amplifier 402 that amplifies the signal to a high-voltage amplitude; a logic circuit 403 that controls the opening and closing of the switching element 401 based on information of a set On time Ton and a sub-origin signal ZP1; and a gain adjuster 404 that adjusts the amplification factor (gain) of the amplifier 402 to adjust the amplitude of the high-voltage high-frequency signal applied to the selection optical element AOM1.

If the amplitude of the high-voltage, high-frequency signal applied to the selective optical element AOM1 changes within an allowable range, the diffraction efficiency of the selective optical element AOM1 can be fine-tuned, thereby changing the intensity of the deflected and emitted light beam LB1 (primary diffracted light). Therefore, in this fifth modification, the gain adjuster 404 is adjusted so that the amplitude of the high-voltage, high-frequency signal applied to each selective optical element AOMn increases in order from the driver circuit DRV1 of the selective optical element AOM1 on the side closest to the light source device 14' to the driver circuit DRV6 of the selective optical element AOM6 on the side farther from the light source device 14'. For example, the amplitude of the high-voltage, high-frequency signal applied to the selective optical element AOM6 at the end of the optical path of the light beam LB is set to a value Va6 at which the diffraction efficiency is maximized, while the amplitude of the high-voltage, high-frequency signal applied to the selective optical element AOM1 at the beginning of the optical path of the light beam LB is set to a value Va1 at which the diffraction efficiency is reduced within the allowable range. Amplitudes Va2 to Va5 of the high-voltage, high-frequency signals applied to the selective optical elements AOM2 to AOM5 therebetween are set to satisfy Va1 < Va2 < Va3 < Va4 < Va5 < Va6 .

The above settings can mitigate or suppress variations in the intensity of the light beams LB1 to LB6 emitted from the six selection optical elements AOM1 to AOM6. This can suppress variations in the exposure of the patterns drawn by the drawing lines SL1 to SL6, enabling high-precision pattern drawing. Furthermore, the amplitudes Va1 to Va6 of the high-voltage, high-frequency signals set by the driver circuits DRV1 to DRV6 do not need to be increased sequentially; for example, the relationship Va1 = Va2 < Va3 = Va4 < Va5 = Va6 can be achieved. Furthermore, in addition to the method of Modification 5, the intensity of the light beams LB1 to LB6 used to draw the point light SP can be adjusted for each scanning unit U1 to U6 by providing a neutral density filter (ND filter) having a predetermined transmittance in the optical path within each scanning unit U1 to U6.

Furthermore, in the driver circuit DRVn of FIG44 , a switching element 401 switches whether the high-frequency signal from the high-frequency transmission source 400 is transmitted to the amplifier 402. However, to improve the responsiveness (rise characteristics) when switching the selection optical element AOMn on and off, a low-level high-frequency signal, such as one with an intensity of primary diffracted light of 1/1000 or less relative to the intensity when the optical element is on, may be continuously applied to the selection optical element AOMn when the diffraction efficiency is essentially considered to be zero, and an appropriately high-level high-frequency signal may be applied to the selection optical element AOMn only when the optical element is on. FIG45 illustrates the structure of such a driver circuit DRVn, with the structure of the driver circuit DRV1 being representatively shown here, and components identical to those in FIG44 are denoted by the same reference numerals.

The configuration of FIG45 adds two resistors, RE1 and RE2, connected in series. The series circuit of resistors RE1 and RE2 is inserted into the high-frequency transmission source 400, in parallel with the switching element 401, just before the switching element 401. A high-frequency signal from the high-frequency transmission source 400, divided by the resistance ratio RE2/(RE1+RE2), is constantly applied to the amplifier 402. When resistor RE2 is configured as a variable resistor and the switching element 401 is in the off (non-conducting) state, the level of the high-frequency signal applied to the selection optical element AOM1 is adjusted so that the intensity of the primary diffracted light emitted from the selection optical element AOM1, i.e., the light beam LB1, is sufficiently low (e.g., less than 1/1000 of its original intensity). In this way, the high-frequency signal is biased (increased) to the selection optical element AOM1 via resistors RE1 and RE2, thereby improving responsiveness. In addition, in this case, the intensity is also extremely weak during the period when the switching element 401 is in the Off (non-conducting) state, but since the light beam LB1 is incident on the corresponding scanning unit U1, when the conveying speed of the substrate FS is reduced or stopped during the drawing action due to some fault, the light gate located at the exit of the light source device 14' (14A', 14B') is closed, or a dimming filter is inserted.

[Variant 6]

In each of the above embodiments and modifications, a pattern is drawn along a drawing line SLn on the surface of the substrate FS curved into a cylindrical surface by a plurality of scanning units Un, with the sheet-like substrate FS in close contact with the outer circumferential surface of the rotating drum DR. However, a structure may also be adopted in which the substrate FS is supported in a planar shape and the exposure process is performed while being transported in the longitudinal direction, as disclosed in, for example, International Publication No. 2013/150677. In this case, if the surface of the substrate FS is set to be parallel to the XY plane, for example, the plurality of scanning units U1 to U6 can be arranged in such a manner that the irradiation center axes Le1, Le3, Le5 of the odd-numbered scanning units U1, U3, U5 and the irradiation center axes Le2, Le4, Le6 of the even-numbered scanning units U2, U4, U6 shown in Figures 23 and 24 are parallel to the Z axis when viewed in a plane parallel to the XZ plane and are located in the X direction at a certain interval.

Claims (23)

1. A pattern drawing apparatus that projects a scanning point of a light beam whose intensity is modulated according to drawing data onto an irradiated object and draws a pattern on the irradiated object, characterized in that it comprises: The drawing head is equipped with multiple drawing units that scan different areas of the irradiated object with the scanning points. The multiple drawing units generate the scanning points through a rotating multifaceted mirror and optical lens system that deflects the light beam. A rotation control system that synchronizes the rotational polyhedrons of the plurality of drawing units in a manner that makes the scanning periods of the scanning points based on the plurality of drawing units different from each other. An origin sensor, which is located in each of the plurality of drawing units, outputs a pulse-shaped origin signal indicating the timing of the start of scanning at the scanning point for each reflecting surface of the rotating multifaceted mirror; A light source device includes an optical modulator, an optical fiber amplifier, and a wavelength conversion element, wherein the optical modulator switches and outputs one of a rapidly rising and falling pulsed first type of light and a moderately rising pulsed second type of light generated at the same frequency based on the depicted data; the optical fiber amplifier receives either the first type of light or the second type of light output from the optical modulator; and the wavelength conversion element converts the first type of light amplified by the optical fiber amplifier into an ultraviolet band and outputs it as the light beam. Multiple selective optical elements are arranged corresponding to each of the multiple drawing units in such a way that the light beam from the light source device passes in a straight line, and the light beam is selectively deflected toward one of the multiple drawing units by electrical control; and The drawing control unit, in response to the origin signal generated from a specific drawing unit among the plurality of drawing units, controls the beam to deflect toward the specific drawing unit by a corresponding selective optical element among the plurality of selective optical elements, and controls the drawing data corresponding to the pattern to be drawn by the specific drawing unit to be applied to the light modulator. 2. The pattern drawing device as claimed in claim 1, characterized in that, The number of each of the plurality of drawing units and the plurality of selectable optical elements is set according to the scanning efficiency, wherein the scanning efficiency is determined by the relationship between the rotation angle β corresponding to one of the reflecting surfaces of the rotating polygonal mirror and the rotation angle α of the rotating polygonal mirror that enables the light beam to be incident on the optical lens system. 3. The pattern drawing device as claimed in claim 2, characterized in that, The scanning efficiency is set to 1/3, and the number of each of the plurality of drawing units and the plurality of selectable optical elements is set to 3. 4. The pattern drawing device according to any one of claims 1 to 3, characterized in that, A lens is provided that is disposed between each of the plurality of selectable optical elements and such that the plurality of selectable optical elements are respectively positioned as focal points of each other. 5. The pattern drawing device as claimed in claim 4, characterized in that, It also includes a condenser lens and a collimating lens, which are disposed in the optical path between the first selective optical element closest to the light source device and the light source device among the plurality of selective optical elements. The condenser lens converges the light beam from the light source device to a focal position, and the collimating lens incident on the light beam from the condenser lens and makes the light beam parallel to a predetermined beam diameter. 6. The pattern drawing device as claimed in claim 5, characterized in that, The rotation control system includes: A timing measurement unit measures correction information corresponding to the error in the relative phase difference of multiple origin signals output from the origin sensors of each of the multiple drawing units, in order to synchronize the rotation of the rotating polyhedrons of each of the multiple drawing units; and The servo control unit controls each motor that drives the rotation of the plurality of rotating polyhedrons based on the correction information. 7. The pattern drawing device as claimed in claim 6, characterized in that, When the multiple origin signals are generated stably with a predetermined time difference, the timing measurement unit causes the multiple selection optical elements to output a drawing enable signal for selectively performing the deflection operation of the light beam based on the origin signals. 8. The pattern drawing apparatus as claimed in claim 7, characterized in that, The plurality of selectable optical elements are acousto-optic modulation elements that are transmissive relative to the beam and change the direction of travel of the incident beam by diffraction in response to the depiction enable signal. 9. The pattern drawing device as claimed in claim 4, characterized in that, The light source device includes: The first solid-state laser element generates the first type of light in a pulsed manner in response to a clock pulse of a specified frequency; A second solid-state laser element, which generates the second type of light in a pulsed manner in response to the clock pulse; and The first polarizing beam splitter directs the first type of light and the second type of light toward the optical modulator through transmission and reflection. 10. The pattern drawing apparatus as claimed in claim 9, characterized in that, The energy of the pulsed first type of light incident on the light modulator is approximately the same as the energy of the pulsed second type of light. 11. The pattern drawing apparatus as claimed in claim 10, characterized in that, The depiction data is bitmap data representing on and off states using binary data. In the on state, the scan point is illuminated on the pixel corresponding to a point of the pattern to be depicted; in the off state, the scan point is not illuminated on the pixel. The optical modulator includes an electro-optical element that switches the polarization direction of the first type of light and the second type of light incident via the first polarizing beam splitter according to the value of the binary data of the pixel in the bitmap data. 12. The pattern drawing apparatus as claimed in claim 11, characterized in that, The optical modulator includes a second polarization beam splitter located after the electro-optical element. When the second polarizing beam splitter is in the on state of illuminating the scanning point of the pixel, it guides the first type of light after passing through the electro-optical element to the fiber amplifier. When it is in the off state of not illuminating the scanning point of the pixel, it guides the second type of light after passing through the electro-optical element to the fiber amplifier. 13. The pattern drawing apparatus as claimed in claim 12, characterized in that, When the frequency of the clock pulse is set to Fs, the size of the scanning point on the irradiated object is set to Ds, and the scanning speed of the scanning point on the irradiated object based on the rotating polyhedron is set to Vs, the relationship is set to Fs≥Vs/Ds. 14. The pattern drawing apparatus as claimed in any one of claims 1 to 3, characterized in that, The light source device includes: The first solid-state laser element generates the first type of light in a pulsed manner in response to a clock pulse of a specified frequency; A second solid-state laser element, which generates the second type of light in a pulsed manner in response to the clock pulse; and The first polarizing beam splitter directs the first type of light and the second type of light toward the optical modulator through transmission and reflection. 15. The pattern drawing apparatus as claimed in claim 14, characterized in that, The energy of the pulsed first type of light incident on the light modulator is approximately the same as the energy of the pulsed second type of light. 16. The pattern drawing apparatus as claimed in claim 15, characterized in that, The depiction data is bitmap data representing on and off states using binary data. In the on state, the scan point is illuminated on the pixel corresponding to a point of the pattern to be depicted; in the off state, the scan point is not illuminated on the pixel. The optical modulator includes an electro-optical element that switches the polarization direction of the first type of light and the second type of light incident via the first polarizing beam splitter according to the value of the binary data of the pixel in the bitmap data. 17. The pattern drawing apparatus as claimed in claim 16, characterized in that, The optical modulator includes a second polarization beam splitter located after the electro-optical element. When the second polarizing beam splitter is in the on state of illuminating the scanning point of the pixel, it guides the first type of light after passing through the electro-optical element to the fiber amplifier. When it is in the off state of not illuminating the scanning point of the pixel, it guides the second type of light after passing through the electro-optical element to the fiber amplifier. 18. The pattern drawing apparatus as claimed in claim 17, characterized in that, When the frequency of the clock pulse is set to Fs, the size of the scanning point on the irradiated object is set to Ds, and the scanning speed of the scanning point on the irradiated object based on the rotating polyhedron is set to Vs, the relationship is set to Fs≥Vs/Ds. 19. The pattern drawing apparatus as claimed in claim 5, characterized in that, The light source device includes: The first solid-state laser element generates the first type of light in a pulsed manner in response to a clock pulse of a specified frequency; A second solid-state laser element, which generates the second type of light in a pulsed manner in response to the clock pulse; and The first polarizing beam splitter directs the first type of light and the second type of light toward the optical modulator through transmission and reflection. 20. The pattern drawing apparatus as claimed in claim 19, characterized in that, The energy of the pulsed first type of light incident on the light modulator is approximately the same as the energy of the pulsed second type of light. 21. The pattern drawing apparatus as claimed in claim 20, characterized in that, The depiction data is bitmap data representing on and off states using binary data. In the on state, the scan point is illuminated on the pixel corresponding to a point of the pattern to be depicted; in the off state, the scan point is not illuminated on the pixel. The optical modulator includes an electro-optical element that switches the polarization direction of the first type of light and the second type of light incident via the first polarizing beam splitter according to the value of the binary data of the pixel in the bitmap data. 22. The pattern drawing apparatus as claimed in claim 21, characterized in that, The optical modulator includes a second polarization beam splitter located after the electro-optical element. When the second polarizing beam splitter is in the on state of illuminating the scanning point of the pixel, it guides the first type of light after passing through the electro-optical element to the fiber amplifier. When it is in the off state of not illuminating the scanning point of the pixel, it guides the second type of light after passing through the electro-optical element to the fiber amplifier. 23. The pattern drawing apparatus as claimed in claim 22, characterized in that, When the frequency of the clock pulse is set to Fs, the size of the scanning point on the irradiated object is set to Ds, and the scanning speed of the scanning point on the irradiated object based on the rotating polyhedron is set to Vs, the relationship is set to Fs≥Vs/Ds.